The Blue Dot Issue 9: Artificial Intelligence and the Future of Education

The ninth issue of The Blue DOT takes on artificial intelligence and the future of education. Artificial Intelligence (AI) has caught the imagination of the world and has been gaining popularity in the business sector. Rather than replacing teachers and making learning impersonal, AI could take learning to a completely new level. Read what academicians, policymakers, practitioners and researchers have to say about the use of Artificial Intelligence (AI) in Education.

The ninth issue of The Blue DOT takes on artificial intelligence and the future of education. Artificial Intelligence (AI) has caught the imagination of the world and has been gaining popularity in the business sector. Rather than replacing teachers and making learning impersonal, AI could take learning to a completely new level. Read what academicians, policymakers, practitioners and researchers have to say about the use of Artificial Intelligence (AI) in Education.


You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.


ISSUE 9, 2018<br />


Director<br />


United Nations <strong>Education</strong>al, Scientific <strong>and</strong> Cultural<br />

Organization | Mahatama G<strong>and</strong>hi Institute <strong>of</strong><br />

<strong>Education</strong> for Peace <strong>and</strong> Sustainable Development<br />

35 Ferozshah Road, ICSSR, Building, 1 st Floor<br />

New Delhi 110001, INDIA<br />

THE BLUE DOT features articles showcasing UNESCO<br />

MGIEP’s activities <strong>and</strong> areas <strong>of</strong> interest. <strong>The</strong> magazine’s<br />

overarching <strong>the</strong>me is <strong>the</strong> relationship between education, peace,<br />

sustainable development <strong>and</strong> global citizenship. THE BLUE<br />

DOT’s role is to engage with readers on <strong>the</strong>se issues in a fun<br />

<strong>and</strong> interactive manner. <strong>The</strong> magazine is designed to address<br />

audiences across generations <strong>and</strong> walks <strong>of</strong> life, <strong>the</strong>reby taking <strong>the</strong><br />

discourse on education for peace, sustainable development <strong>and</strong><br />

global citizenship beyond academia, civil society organisations<br />

<strong>and</strong> governments, to <strong>the</strong> actual stakeholders.<br />


Programme Specialist - Science <strong>of</strong> Learning<br />

THE BLUE DOT is published biannually.<br />


<strong>The</strong> <strong>Blue</strong> <strong>Dot</strong> is available free <strong>of</strong> charge.<br />

To receive all future issues <strong>of</strong> <strong>the</strong> THE BLUE DOT,<br />

subscribe to bluedot.mgiep@unesco.org<br />


Akriti Mehra, UNESCO MGIEP<br />

DESIGN<br />

Prasun Mazumdar Design | www.pmdindia.in<br />


Communications Specialist<br />


Disclaimer: <strong>The</strong> views <strong>and</strong> opinions expressed in this magazine<br />

do not necessarily reflect <strong>the</strong> <strong>of</strong>ficial policy or position <strong>of</strong><br />


<strong>The</strong> image used on <strong>the</strong> cover <strong>of</strong> this issue <strong>of</strong> <strong>The</strong> <strong>Blue</strong> <strong>Dot</strong> is<br />

purely representational <strong>and</strong> conceptual in nature.

“Look again at that dot.<br />

That’s here. That’s home. That’s us.<br />

On it, everyone you love,<br />

everyone you know, everyone you ever heard <strong>of</strong>,<br />

every human being who ever was,<br />

lived out <strong>the</strong>ir lives.<br />

<strong>The</strong> aggregate <strong>of</strong> our joy <strong>and</strong> suffering<br />

thous<strong>and</strong>s <strong>of</strong> confident religions,<br />

ideologies, <strong>and</strong> economic doctrines,<br />

every hunter <strong>and</strong> forager, every hero <strong>and</strong> coward,<br />

every creator <strong>and</strong> destroyer <strong>of</strong> civilization,<br />

every king <strong>and</strong> peasant, every young couple in love,<br />

every mo<strong>the</strong>r <strong>and</strong> fa<strong>the</strong>r, hopeful child,<br />

inventor <strong>and</strong> explorer, every teacher <strong>of</strong> morals,<br />

every corrupt politician, every superstar,<br />

every supreme leader, every saint<br />

<strong>and</strong> sinner in <strong>the</strong> history <strong>of</strong> our species lived <strong>the</strong>reon<br />

a mote <strong>of</strong> dust suspended in a sunbeam.”<br />



director’s<br />

message<br />

A<br />

child somewhere<br />

in “poor” India is<br />

having trouble with<br />

introductory algebra.<br />

This is new content for <strong>the</strong> child <strong>and</strong> <strong>the</strong><br />

bigger problem <strong>the</strong> child faces is that <strong>the</strong><br />

teacher is hardly ever in school. <strong>The</strong>re is<br />

some hope, however, as s/he has good<br />

access to <strong>the</strong> Internet <strong>and</strong> to <strong>the</strong> many<br />

online teaching tools on <strong>the</strong> subject.<br />

Despite this, <strong>the</strong> child cannot afford a<br />

personalised online tutor <strong>and</strong> continues<br />

to struggle to underst<strong>and</strong> <strong>the</strong> basic<br />

concepts. That, I believe, might be a<br />

close description <strong>of</strong> <strong>the</strong> current education<br />

system in India <strong>and</strong> many parts <strong>of</strong> <strong>the</strong><br />

world.’<br />

Now visualise a system where <strong>the</strong><br />

child is given access to an online global<br />

intelligent education platform. As part<br />

<strong>of</strong> this system, <strong>the</strong> child is paired with<br />

a personalised BOT right after s/he<br />

enters school. This BOT develops a keen<br />

underst<strong>and</strong>ing <strong>of</strong> <strong>the</strong> child’s attributes <strong>and</strong><br />

learning preferences by evaluating data <strong>of</strong><br />

<strong>the</strong> child’s ongoing learning experiences.’<br />

<strong>The</strong> BOT has access to a finite set<br />

<strong>of</strong> possible interventions – arising from<br />

learner-centric data derived from <strong>the</strong><br />

experiences <strong>of</strong> millions <strong>of</strong> children<br />

learning algebra worldwide – to help <strong>the</strong><br />

child overcome learning problems. <strong>The</strong><br />

BOT can parse all <strong>the</strong> learner-centric data<br />

to identify specific interventions suited to<br />

<strong>the</strong> particular child who is its responsibility<br />

– functioning as a virtual, personalised<br />

<strong>Artificial</strong> <strong>Intelligence</strong><br />

for <strong>Education</strong><br />

tutor. Personalised learning, <strong>the</strong> holy grail<br />

<strong>of</strong> education, is a definite reality in this<br />

hypo<strong>the</strong>tical scenario. As I perceive it,<br />

making this a reality for children today is a<br />

distinct possibility.<br />

Is this just in <strong>the</strong> realms <strong>of</strong> fiction or<br />

a potential reality? <strong>The</strong> recent growing<br />

importance <strong>of</strong> <strong>Artificial</strong> <strong>Intelligence</strong> (AI)<br />

seems to point such a scenario as a<br />

reality within our reach. However, <strong>the</strong>re<br />

has also been growing concern over <strong>the</strong><br />

ethical <strong>and</strong> moral use <strong>of</strong> AI coupled<br />

with a fear <strong>of</strong> AI <strong>and</strong> making many jobs<br />

redundant. Although <strong>the</strong> use <strong>of</strong> AI in <strong>the</strong><br />

education sector is at <strong>the</strong> nascent stage,<br />

<strong>the</strong>se fears are reflected as well <strong>and</strong> in<br />

particular to <strong>the</strong> rising sentiment that AI<br />

will replace teachers.<br />

Doucher <strong>and</strong> Evers, in <strong>the</strong>ir seminal<br />

book, “Teaching in <strong>the</strong> Fourth Revolution:<br />

St<strong>and</strong>ing at <strong>the</strong> Precipice”, argue that AI<br />

will no doubt have a major impact in <strong>the</strong><br />

way learning <strong>and</strong> teaching will happen in<br />

<strong>the</strong> future. <strong>The</strong>re is no looking back.<br />

But what that future will look like is for<br />

us here at <strong>the</strong> present to guide that journey.<br />

One thing is for sure: <strong>the</strong> traditional roles<br />

<strong>of</strong> learners <strong>and</strong> teachers will evolve <strong>and</strong><br />

will need to be more symbiotic.<br />

But before we go into details <strong>of</strong> AI’s role<br />

in education, some underst<strong>and</strong>ing <strong>of</strong> its<br />

scope is necessary. AI experts categorise<br />

three main types <strong>of</strong> AI: Strong AI, Weak<br />

AI <strong>and</strong> <strong>Artificial</strong> General <strong>Intelligence</strong>.<br />

<strong>The</strong>se are well explored <strong>and</strong> <strong>the</strong> myth<br />

behind AI is discussed in this issue <strong>of</strong><br />

<strong>The</strong> <strong>Blue</strong> <strong>Dot</strong>. <strong>The</strong>se systems have been<br />

elaborated upon in detail <strong>and</strong> <strong>the</strong> myths<br />

surrounding <strong>the</strong>m have been dispelled by<br />

some <strong>of</strong> <strong>the</strong> world’s foremost AI experts.<br />

<strong>The</strong> second part <strong>of</strong> <strong>the</strong> AI <strong>and</strong><br />

education relationship explores how<br />

AI can be used to improve learning<br />

<strong>and</strong> make it relevant for 21 st century<br />

challenges. Articles written on this topic<br />

will explore whe<strong>the</strong>r <strong>the</strong> new world that<br />

I had proposed in <strong>the</strong> beginning <strong>of</strong> my<br />

message is achievable <strong>and</strong> if yes, what<br />

needs to be done to achieve that holy grail<br />

<strong>of</strong> education.<br />

This issue was based on one <strong>of</strong> <strong>the</strong> five<br />

main <strong>the</strong>mes <strong>of</strong> our recently concluded<br />

TECH 2018 conference in Vizag, India,<br />

titled ‘AI <strong>and</strong> <strong>Education</strong>’, for which<br />

some <strong>of</strong> <strong>the</strong> world’s leading experts were<br />

invited to elaborate on <strong>the</strong> <strong>the</strong>me. Equally<br />

important to have is <strong>the</strong> discussion on <strong>the</strong><br />

privacy <strong>and</strong> ownership <strong>of</strong> information that<br />

will be generated by learners <strong>and</strong> teachers.<br />

<strong>The</strong> use <strong>of</strong> information is a contentious<br />

issue that needs to be addressed early as we<br />

embark on <strong>the</strong> exciting <strong>and</strong> path-breaking<br />

journey that I have termed <strong>the</strong> <strong>The</strong><br />

Internet <strong>of</strong> <strong>Education</strong> (TIE). I hope that<br />

this journey will culminate in education<br />

<strong>and</strong> knowledge becoming accessible to<br />

<strong>and</strong> beneficial for all <strong>of</strong> humanity. A High<br />

Level Policy Forum on development <strong>of</strong><br />

guidelines was initiated at <strong>the</strong> TECH<br />

2018, after which <strong>the</strong> Vizag Declaration<br />

was adopted at <strong>the</strong> closing ceremony <strong>of</strong><br />

<strong>the</strong> conference, calling for development <strong>of</strong><br />

Guidelines for Digital Learning.<br />

We here at UNESCO MGIEP have<br />

already begun this journey by partnering<br />

with like-minded organisations <strong>and</strong> have<br />

developed a prototype platform called<br />

CHI (Collective Human <strong>Intelligence</strong>). This<br />

platform allows educators <strong>and</strong> learners<br />

to develop curriculum, lesson plans <strong>and</strong><br />

assessments in an interactive, immersive<br />

<strong>and</strong> experiential environment, which is<br />

supported by AI that is able to provide<br />

feedback to students <strong>and</strong> educators <strong>of</strong><br />

progress <strong>and</strong> suggestions to improve<br />

learning.<br />


Director, UNESCO MGIEP<br />


0 1


<strong>Artificial</strong> <strong>Intelligence</strong> &<br />

<strong>Education</strong> & UNESCO’s<br />

Way Forward<br />

Currently, <strong>the</strong>re is no universally<br />

recognised institutional framework<br />

on <strong>the</strong> ethics <strong>of</strong> AI, particularly in<br />

relation to education.<br />




T<br />

he permeation <strong>of</strong> new technologies <strong>and</strong> <strong>Artificial</strong><br />

<strong>Intelligence</strong> (AI) is changing how we teach, live,<br />

work <strong>and</strong> learn. Undoubtedly, AI has <strong>the</strong> potential<br />

to widely benefit us <strong>and</strong> resolve various societal<br />

issues. At <strong>the</strong> same time, it also presents certain challenges that<br />

need to be addressed. Yet, if we are able to prepare ourselves to<br />

tackle <strong>the</strong>se challenges, we will be able to benefit widely from this<br />

collective intelligence <strong>and</strong> <strong>the</strong> ‘Fourth Industrial Revolution’.<br />

In <strong>the</strong> field <strong>of</strong> education, AI is expected to have a<br />

significant impact, transforming <strong>the</strong> way teachers work<br />

by <strong>of</strong>fering new teaching aids. It could revolutionise <strong>the</strong> way<br />

in which students learn through personalised learning <strong>and</strong> through<br />

greater access to knowledge thus also potentially facilitating more<br />

inclusive education.<br />

am very pleased to share that our Category I Institute in<br />

<strong>the</strong> Asia-Pacific, <strong>the</strong> UNESCO Mahatma G<strong>and</strong>hi Institute<br />

<strong>of</strong> <strong>Education</strong> for Peace <strong>and</strong> Sustainable Development<br />

(MGIEP), has taken <strong>the</strong> lead on SEL <strong>and</strong> exploring ways<br />

AI can be harnessed to integrate SEL into our <strong>Education</strong><br />

systems.<br />

I am particularly pleased with <strong>the</strong> progress that has been made with<br />

<strong>the</strong> development <strong>of</strong> an open-source indigenously-designed <strong>and</strong><br />

collaborative curriculum development <strong>and</strong> learning platform that<br />

allows educators <strong>and</strong> students to innovate, develop <strong>and</strong> implement<br />

curriculum. This Collective Human <strong>Intelligence</strong> platform will use<br />

advanced AI <strong>and</strong> data analytics tools to personalise <strong>the</strong> learning<br />

experience for each user, as well as address data ownership, privacy<br />

<strong>and</strong> security issues.<br />

Clearly, <strong>the</strong> possibilities are tremendous. However, <strong>the</strong>re are some<br />

associated risks, particularly those related to individual privacy<br />

<strong>and</strong> ethics. UNESCO, by its m<strong>and</strong>ate, is ideally placed to<br />

play a key role by supporting Member States to increase<br />

accessibility <strong>of</strong> technology, promoting <strong>the</strong> use <strong>of</strong> opensource<br />

AI platforms <strong>and</strong> tools that encourage learning.<br />

Additionally, UNESCO’s role will be crucial to rethink educational<br />

curricula, by bringing into focus social <strong>and</strong> emotional learning to<br />

build <strong>the</strong> emotional intelligence <strong>of</strong> future generations <strong>of</strong> learners.<br />

Currently, <strong>the</strong>re is no universally recognised institutional framework<br />

on <strong>the</strong> ethics <strong>of</strong> AI, particularly in relation to education. It is<br />

UNESCO’s duty, with <strong>the</strong> support <strong>of</strong> Member States,<br />

to lead a global debate in order to establish a globally<br />

accepted framework <strong>of</strong> ethical principles <strong>and</strong> we have<br />

already begun work in this area.<br />

Fur<strong>the</strong>rmore, I welcome <strong>the</strong> collaboration between <strong>the</strong> Institute<br />

<strong>and</strong> <strong>the</strong> State Government <strong>of</strong> Andhra Pradesh, India, as well<br />

as o<strong>the</strong>r partners <strong>and</strong> experts in <strong>the</strong> field <strong>of</strong> EdTech to establish<br />

global guidelines for <strong>the</strong> exponentially growing field <strong>of</strong> digital<br />

learning tools <strong>and</strong> methodologies. <strong>The</strong>se guidelines will help <strong>the</strong><br />

development <strong>of</strong> digital pedagogical tools to effectively harness this<br />

new technology for <strong>the</strong> betterment <strong>of</strong> society. I also look forward<br />

to <strong>the</strong> outcomes <strong>of</strong> <strong>the</strong> Institute’s annual international EdTech<br />

conference, TECH 2018, for which AI <strong>and</strong> <strong>Education</strong> was a<br />

key focus area. I wish <strong>the</strong> Institute all <strong>the</strong> best in its endeavours<br />

to mainstream SEL through <strong>the</strong> use <strong>of</strong> AI: a vital issue as we<br />

deliberate on <strong>the</strong> ethics <strong>of</strong> AI.<br />

<strong>The</strong> ninth issue <strong>of</strong> <strong>The</strong> <strong>Blue</strong> <strong>Dot</strong>, focuses on AI <strong>and</strong> <strong>the</strong> future <strong>of</strong><br />

education. It explores debates on <strong>the</strong> use <strong>of</strong> AI in education <strong>and</strong> <strong>the</strong><br />

potential it <strong>of</strong>fers to complement existing education systems <strong>and</strong><br />

teachers, as well as in facilitating personalised learning. Indeed, I<br />

ISSUE • 9<br />

0 3


Approach to<br />

<strong>Artificial</strong> <strong>Intelligence</strong> for<br />

<strong>Education</strong> <strong>and</strong> Training<br />

AI is concerned with underst<strong>and</strong>ing<br />

intelligence, by investigating <strong>the</strong><br />

design <strong>of</strong> machines that have<br />

intelligent abilities.<br />


A<br />

rtificial <strong>Intelligence</strong> (AI) is no<br />

longer just a Hollywood invention,<br />

it is already impacting our lives.<br />

From truck drivers, whose role may be<br />

replaced by autonomous vehicles, to paralegal workers who can<br />

be outsmarted by AI s<strong>of</strong>tware, <strong>the</strong> world <strong>of</strong> work is changing with<br />

increasing pace. <strong>The</strong> implications for education <strong>and</strong> training are<br />

significant <strong>and</strong> far reaching. We need to prepare people for a future<br />

that will likely be evolving quickly, <strong>and</strong> not everyone will have a<br />

positive experience. However, it is not all doom <strong>and</strong> gloom. We<br />

can also reap huge benefits from AI if we approach it<br />

intelligently.<br />

to s<strong>of</strong>tware that can support students with individual tutoring for<br />

<strong>the</strong>ir school <strong>and</strong> university studies, <strong>and</strong> learning platforms that<br />

provide educational support informed by neuroscience. We can<br />

track students’ interactions, from every mouse movement <strong>and</strong> each<br />

keystroke <strong>and</strong> use AI to analyse each student’s performance in<br />

detail. This analysis provides personalised learning for <strong>the</strong> student<br />

<strong>and</strong> detailed feedback about <strong>the</strong> learning status <strong>of</strong> every student in<br />

<strong>the</strong>ir class for teachers. Well-designed AI can help us tackle some<br />

<strong>of</strong> our toughest educational challenges, such as global teacher<br />

shortages <strong>and</strong> <strong>the</strong> achievement gap between <strong>the</strong> more <strong>and</strong> less<br />

advantaged members <strong>of</strong> <strong>the</strong> population.<br />

But what exactly is AI, <strong>and</strong> how can we approach it intelligently?<br />


PROFESSOR,UCL Knowledge Lab<br />

AI is concerned with underst<strong>and</strong>ing intelligence by<br />

investigating <strong>the</strong> design <strong>of</strong> machines that have intelligent<br />

abilities. AI has gained traction recently because <strong>of</strong> <strong>the</strong> ‘perfect<br />

storm’ brought about by <strong>the</strong> availability <strong>of</strong> large quantities <strong>of</strong> data,<br />

powerful computer processing with large clouds <strong>of</strong> cheap, accessible<br />

storage <strong>and</strong> AI algorithms that can learn. <strong>The</strong>se 3 factors: data,<br />

computing (power <strong>and</strong> storage), <strong>and</strong> sophisticated AI means that we<br />

can build vast AI ‘knowledge’ bases that can be sifted <strong>and</strong> searched<br />

with stunning accuracy to pinpoint <strong>the</strong> minutiae <strong>of</strong> detail required<br />

to diagnose a disease, process an image, beat world champion game<br />

players <strong>and</strong> to help us navigate through our complex physical <strong>and</strong><br />

virtual worlds.<br />

<strong>The</strong> intelligent way to approach AI is to focus on planning<br />

educational <strong>and</strong> training systems that use AI wisely to<br />

prepare people to work <strong>and</strong> live with AI. <strong>The</strong> situation is<br />

summarised in Figure 1<br />

This figure illustrates that we need to:<br />

1. Engage educators <strong>and</strong> trainers in working with AI developers to<br />

build AI technology that is specifically designed for education <strong>and</strong><br />

training. This AI development has already started: from experiential<br />

digital learning driven by AI powered virtual role play simulations,<br />

2. Educate people about AI, so that <strong>the</strong>y can use it successfully,<br />

ethically <strong>and</strong> knowledgeably. For some people, this underst<strong>and</strong>ing<br />

about AI needs to be more extensive, to train people to: make<br />

responsible ethical decisions about how <strong>and</strong> when AI should be<br />

used; or develop <strong>the</strong> next generation <strong>of</strong> AI technology. Educating<br />

people about AI, can also be assisted by well-designed AI systems.<br />

3. Prioritise <strong>the</strong> development <strong>of</strong> our own human intelligence so that<br />

we focus attention on <strong>the</strong> elements <strong>of</strong> intelligence that are not (yet)<br />

automatable with AI, such as social intelligence <strong>and</strong> metacognitive<br />

intelligence. This too can be assisted by <strong>the</strong> smart use <strong>of</strong> welldesigned<br />

AI to track <strong>and</strong> support our learning <strong>and</strong> increase our<br />

intelligence. For more detail about this, please see my 2018 book:<br />

Machine Learning <strong>and</strong> Human <strong>Intelligence</strong>.<br />

<strong>The</strong> thoughtful design <strong>of</strong> AI approaches to education <strong>and</strong><br />

training has <strong>the</strong> potential to provide significant benefits<br />

to educators, learners, parents <strong>and</strong> managers. It is a reason<br />

for great excitement <strong>and</strong> concerted intelligent effort.<br />

ISSUE • 9<br />

0 5


ISSUE<br />

36<br />

44<br />

9, 2018<br />

Director’s Message<br />

<strong>Artificial</strong> <strong>Intelligence</strong> for <strong>Education</strong><br />

By Anantha Kumar Duraiappah | Director, UNESCO MGIEP<br />

Foreword<br />

• <strong>Artificial</strong> <strong>Intelligence</strong> & <strong>Education</strong> & UNESCO’s<br />

Way Forward<br />

By Audrey Azoulay | Director-General, Unesco<br />

01<br />

02<br />

• Approach to <strong>Artificial</strong> <strong>Intelligence</strong> for <strong>Education</strong> &<br />

Training<br />

By Rose Luckin | Pr<strong>of</strong>essor, UCL Knowledge Lab<br />

• Declaration <strong>of</strong> Independence II<br />

By Dan Shefet<br />

• Dreams & Reality: How AI will change <strong>Education</strong><br />

By Harri Ketamo<br />

30<br />


• CHI – a Knowledge Sharing Digital Platform Powered<br />

by Machine Learning<br />

Saurabh Roy | Chief Technology Officer, UNESCO MGIEP<br />


40<br />

• Interview with Dr. Avik Sarkar, Head – Data Analytics<br />

Cell at NITI Aayog, Govt. <strong>of</strong> India<br />


• Can <strong>Artificial</strong> <strong>Intelligence</strong> Help Us Achieve Universal<br />

Quality <strong>Education</strong>?<br />

By James Hodson<br />

• <strong>Artificial</strong> <strong>Intelligence</strong> for <strong>Education</strong><br />

By Soma S Dhavala<br />

50<br />

06<br />

42<br />



What we have been up to at UNESCO MGIEP<br />

• <strong>The</strong> Digital Transformation <strong>of</strong> K-12: A Viewpoint<br />

By Cathie Norris & Elliot Soloway<br />

• Peace Machine as a Re-emerging G<strong>and</strong>hi<br />

By Timo Honkela<br />

• Promoting Peace: How Can Data Science Assist?<br />

By Srujana Merugu<br />


• Personalising ‘Learning’ - Can AI Promise<br />

Customised <strong>Education</strong> For ‘Humanity’<br />

N<strong>and</strong>ini Chatterjee Singh & Raunak Jain<br />

TECH 2018<br />

What to Expect at TECH 2018?<br />


Learning with AI<br />


opinion<br />

Expert<br />

Perspectives<br />

<strong>The</strong> Digital Transformation<br />

<strong>of</strong> K-12: A Viewpoint<br />


<strong>Artificial</strong> <strong>Intelligence</strong> <strong>and</strong> <strong>the</strong> <strong>Future</strong> <strong>of</strong><br />

<strong>Education</strong><br />

<strong>Artificial</strong> <strong>Intelligence</strong> (AI) has caught <strong>the</strong><br />

imagination <strong>of</strong> <strong>the</strong> world <strong>and</strong> has been gaining<br />

popularity in <strong>the</strong> business sector. To date, <strong>the</strong><br />

use <strong>of</strong> AI in education has been limited <strong>and</strong> at<br />

times contested; most arguments against it centre<br />

on how it will make teachers redundant <strong>and</strong><br />

learning more automated than it is currently.<br />

Ra<strong>the</strong>r than replacing teachers <strong>and</strong> making<br />

learning impersonal, AI could take learning to a<br />

completely new level. Read what academicians,<br />

policymakers, practitioners <strong>and</strong> researchers have<br />

to say about <strong>the</strong> use <strong>of</strong> <strong>Artificial</strong> <strong>Intelligence</strong><br />

(AI) in <strong>Education</strong><br />

CATHIE NORRIS, Regents Pr<strong>of</strong>essor, Department <strong>of</strong> Learning Technologies, College <strong>of</strong> Information at<br />

<strong>The</strong> University <strong>of</strong> North Texas<br />

ELLIOT SOLOWAY, Arthur F. Thurnau Pr<strong>of</strong>essor, Department <strong>of</strong> Cse, College Of Engineering, at<br />

<strong>The</strong> University Of Michigan<br />

Dr. Cathleen Norris is <strong>the</strong> Regents Pr<strong>of</strong>essor & Chairperson, Department <strong>of</strong><br />

Learning Technologies, College <strong>of</strong> Information, University <strong>of</strong> North Texas. Dr.<br />

Norris’ 14 years in K-12 classrooms – <strong>and</strong> receiving Dallas’ Golden Apple Award<br />

– has shaped her university R&D agenda: developing resources to support<br />

K-12 teachers as <strong>the</strong>y move into 21 st century classrooms. From 1995-2001,<br />

Dr. Norris was President <strong>of</strong> <strong>the</strong> National <strong>Education</strong>al Computing Association<br />

(NECA), <strong>and</strong> led its merger with ISTE, <strong>the</strong> International Society for Technology<br />

in <strong>Education</strong>, creating <strong>the</strong> largest, international organisation for technologyminded<br />

educators in <strong>the</strong> world. Dr. Norris was Co-President/President <strong>of</strong> ISTE<br />

from 2001-2004.<br />

Dr. Elliot Soloway is an Arthur F. Thurnau Pr<strong>of</strong>essor, Department <strong>of</strong> Computer<br />

Science <strong>and</strong> Engineering, College <strong>of</strong> Engineering, with appointments also in<br />

<strong>the</strong> School <strong>of</strong> <strong>Education</strong> <strong>and</strong> School <strong>of</strong> Information, University <strong>of</strong> Michigan. For<br />

15+ years, Dr. Soloway’s R&D efforts have been guided by <strong>the</strong> vision that mobile,<br />

low-cost, networked devices are <strong>the</strong> only way to truly achieve universal 1:1 in<br />

schools – all across <strong>the</strong> globe. In 2001, <strong>the</strong> UMich undergraduates selected him<br />

to receive <strong>the</strong> “Golden Apple Award” as <strong>the</strong> Outst<strong>and</strong>ing Teacher <strong>of</strong> <strong>the</strong> Year at<br />

<strong>the</strong> University <strong>of</strong> Michigan. In 2004 <strong>and</strong> in 2011, College <strong>of</strong> Engineering HKN<br />

Honor Society students selected Dr. Soloway to receive <strong>the</strong> “Distinguished<br />

Teacher <strong>of</strong> <strong>the</strong> Year Award.” In 2001, Norris <strong>and</strong> Soloway founded GoKnow,<br />

Inc., one <strong>of</strong> <strong>the</strong> first mobile learning companies; over 40,000 students<br />

worldwide used GoKnow’s Mobile Learning Environment. In 2012, Norris <strong>and</strong><br />

Soloway founded <strong>the</strong> Intergalactic Mobile Learning Center, <strong>and</strong> with support<br />

from private (e.g., Qualcomm, Google, George Lucas <strong>Education</strong> Foundation)<br />

<strong>and</strong> public (e.g., National Science Foundation) sources, has published <strong>the</strong><br />

Collabrify Suite <strong>of</strong> Apps <strong>and</strong> <strong>the</strong> Collabrify Roadmap System – free, deviceindependent,<br />

tools, used in classrooms all around <strong>the</strong> world.<br />

0 7


<strong>The</strong> Digital Transformation <strong>of</strong> K-12: A<br />

Viewpoint<br />

• “Most 1 school leaders know we’re in <strong>the</strong> midst <strong>of</strong> a digital<br />

transformation, with a size <strong>and</strong> scope larger <strong>and</strong> more significant<br />

than anything we’ve seen before…. Educators know <strong>the</strong>y have to<br />

change but are challenged by <strong>the</strong> convergence <strong>of</strong> technology <strong>and</strong><br />

pedagogy. What technologies can affect learning <strong>and</strong> experience?<br />

How can we get all educators on board? How can we build a case<br />

to support our goals? What’s our future vision?”<br />

<strong>The</strong> above quote sums up <strong>the</strong> status <strong>of</strong> <strong>the</strong> “digital<br />

transformation”, or DX 2 to <strong>the</strong> cognoscenti, in K-12: desire – at<br />

least some – is <strong>the</strong>re, but action – for <strong>the</strong> most part – is lacking.<br />

Arguably, <strong>the</strong> digital transformation <strong>of</strong> K-12 education is <strong>the</strong> most<br />

important task ahead for us in education. It is towards developing<br />

a deeper underst<strong>and</strong>ing <strong>of</strong> what digital transformation is, why it<br />

needs to be undertaken, <strong>and</strong> how it can be accomplished, that we<br />

write this <strong>Blue</strong> <strong>Dot</strong> article. So, let’s begin with <strong>the</strong> “what.”<br />

While those quotes had “business” not<br />

“education” in mind, <strong>the</strong> reality is this:<br />

educating students is as much a business<br />

as banking, retailing, entertaining, etc.<br />

are!<br />

century skills – creativity, communication, collaboration, <strong>and</strong><br />

critical thinking. And, while “driving revenue” is not a focus <strong>of</strong><br />

public education, <strong>the</strong> public purse snaps shut if value is not being<br />

delivered!<br />

• “… using digital technologies” — <strong>The</strong> new acronym for <strong>the</strong><br />

range <strong>of</strong> technologies encompassed by <strong>the</strong> term “digital” is<br />

SMAC 5 : social, mobile, analytics <strong>and</strong> cloud technologies. But<br />

let’s make it “SMAAC” – where <strong>the</strong> second A represents “AI” or<br />

artificial intelligence.<br />


What is Digital transformation?<br />

• “Digital transformation 3 is a fundamental shift in how a firm<br />

delivers value <strong>and</strong> drives revenue.”<br />

• “Digital transformation 4 involves using digital technologies to<br />

retool business models <strong>and</strong> processes to make <strong>the</strong>m more efficient<br />

or effective.”<br />

While those quotes had “business” not “education” in<br />

mind, <strong>the</strong> reality is this: educating students is as much a<br />

business as banking, retailing, entertaining, etc. are!<br />

Let’s unpack those dense sentences:<br />

• “… a fundamental shift …” Not just a “shift” but a “fundamental<br />

shift.” Not digital transition but digital transformation. Words do<br />

count here <strong>and</strong> what is being proposed is truly a dramatic <strong>and</strong><br />

“fundamental” break with <strong>the</strong> past. Analog, paper-<strong>and</strong>-pencil<br />

technologies, while comfortable <strong>and</strong> effective – for some – are not<br />

<strong>the</strong> building blocks <strong>of</strong> <strong>the</strong> future. Digital is based on bits; analog<br />

is based on atoms. Atoms <strong>and</strong> bits are different: duplicating atoms<br />

is expensive; duplicating bits is trivial; changing atoms – erasing a<br />

word in a sentence – is messy <strong>and</strong> time consuming; changing bits –<br />

plopping in a new word into a sentence – is easy, quick, <strong>and</strong> cheap.<br />

Indeed, we are only beginning to appreciate <strong>the</strong> implications <strong>of</strong> a<br />

world based on bits, not atoms.<br />

• “… how a firm delivers value <strong>and</strong> drives revenue.” Delivering<br />

“value” is absolutely <strong>the</strong> top priority in schools. While that value<br />

shares much with <strong>the</strong> past – education must develop students to be<br />

ethical individuals, 21 st century students must develop 21 st<br />

• “… retool business models <strong>and</strong> processes” — Simply put<br />

a “business model” is <strong>the</strong> way a company makes money. For<br />

example, one <strong>of</strong> Apple’s core business models is manufacturing<br />

<strong>and</strong> selling iPhones. But, if <strong>the</strong> rumors 6 are true, Apple is creating<br />

a new streaming service that includes original TV content, so that<br />

collecting subscription fees is an example <strong>of</strong> Apple retooling its<br />

business model <strong>and</strong> processes. In education, ra<strong>the</strong>r than a teacher<br />

st<strong>and</strong>ing in front <strong>of</strong> <strong>the</strong> class <strong>and</strong> delivering a lecture, while<br />

students copy down on paper <strong>the</strong> teacher’s words, schools will need<br />

to develop “digital pedagogies” that exploit SMAAC <strong>and</strong> provide<br />

students with <strong>the</strong> personalised learning experiences that <strong>the</strong>y are<br />

dem<strong>and</strong>ing.<br />

• “… make [<strong>the</strong> business models <strong>and</strong> processes] more efficient or<br />

effective” —<strong>The</strong> terms “efficiency” <strong>and</strong> “effective” are at <strong>the</strong> core<br />

<strong>of</strong> digital transformation. We will explore those terms when we<br />

discuss <strong>the</strong> “how” <strong>of</strong> digital transformation.<br />

Why Is Digital Transformation Needed?<br />

• “Nine 7 out <strong>of</strong> 10 businesses today are undergoing some form<br />

<strong>of</strong> digital transformation <strong>and</strong> customer experience is a stated top<br />

priority.”<br />

While <strong>the</strong>re is debate 8 that <strong>the</strong> 1983 “Nation At Risk” report<br />

may have portrayed American education in an overly poor light,<br />

<strong>the</strong>re is no question that <strong>the</strong>re is considerable room for improving 9<br />

student achievement in U.S. schools. Indeed, in education today,<br />

“customer experience is <strong>the</strong> priority!” <strong>The</strong> following title 10 <strong>of</strong> an<br />

article paints a clear picture:<br />

• “2018 Digital School Districts Survey: Respondents Use<br />

Technology to Put Students First”<br />

<strong>The</strong> survey 11 , tapping into schools in “small, mid-sized <strong>and</strong><br />

large districts in 24 states” conducted by <strong>the</strong> Center for Digital<br />

<strong>Education</strong> found 12 that:<br />

More than nine in 10 respondents to a<br />

survey on <strong>the</strong> topic reported that in <strong>the</strong><br />

pursuit <strong>of</strong> encouraging personalised<br />

learning, <strong>the</strong>ir districts 1) provide s<strong>of</strong>tware<br />

or digital curriculum to classrooms … 2)<br />

provide computing devices to classrooms<br />

• “<strong>The</strong> top learning priority in education for technology use is<br />

personalised learning.<br />

And <strong>the</strong> only way to provide personalised learning is via<br />

technology 13 :<br />

• “More than nine in 10 respondents to a survey on<br />

<strong>the</strong> topic reported that in <strong>the</strong> pursuit <strong>of</strong> encouraging<br />

personalised learning, <strong>the</strong>ir districts 1) provide s<strong>of</strong>tware<br />

or digital curriculum to classrooms … 2) provide<br />

computing devices to classrooms…”<br />

As we will describe in <strong>the</strong> “How” section below, <strong>the</strong>re are several<br />

ways to provide personalised learning – but <strong>the</strong>y all employ<br />

technology.<br />

Besides needing to improve “customer experience,” digital<br />

transformation is needed because those customers – <strong>the</strong> students<br />

<strong>of</strong> today <strong>and</strong> most assuredly <strong>of</strong> tomorrow – are expecting it!<br />

Today’s students are virtually born with a “phone” – <strong>the</strong> students<br />

<strong>of</strong> today do not call <strong>the</strong>ir mobile devices “smartphones” but simply<br />

“phones” – in <strong>the</strong>ir h<strong>and</strong>s. Phones: <strong>The</strong> penetration <strong>of</strong> personally<br />

owning a smartphone is now down 14 to <strong>the</strong> 12-13 year olds in<br />

5 th grade 15 . Yes, today, <strong>the</strong> research 16 is mixed on whe<strong>the</strong>r using<br />

ISSUE • 9<br />

0 9


computers for learning leads to increased student achievement.<br />

But, just as entertainment, medicine, retailing, farming, banking,<br />

etc., etc., are not going back to analog technologies, education is<br />

not going back to paper-<strong>and</strong>-pencil technologies. Educators will<br />

learn how to use digital devices to enhance learning; it’s only a<br />

matter <strong>of</strong> time.<br />

How is Digital Transformation Realized?<br />

In carrying out DX, <strong>the</strong> consensus is that technology itself is<br />

important:<br />

• “Digital transformation is <strong>of</strong>ten driven by <strong>the</strong> IT organization. …<br />

<strong>The</strong> technology is actually <strong>the</strong> center <strong>of</strong> what growth is built on.<br />

<strong>The</strong> technology implications are driving <strong>the</strong> future <strong>of</strong> <strong>the</strong> business<br />

itself.”<br />

And <strong>the</strong> core technology driver will be AI – <strong>Artificial</strong> <strong>Intelligence</strong>:<br />

• “If [your company doesn’t 17 ] have an AI strategy, you’re going to<br />

die in <strong>the</strong> world that’s coming. … more <strong>and</strong> more AI interfaces will<br />

provide natural language inputs <strong>and</strong> outputs, <strong>and</strong> proactively make<br />

decisions on <strong>the</strong> user’s behalf … “<br />

AI – or, more accurately 18 , ML – machine learning 19 – is <strong>the</strong> latest,<br />

cure-all, technological hype in a long line <strong>of</strong> technological hype –<br />

some successful (2G, 3G, 4G), some not (have you seen a 3D TV<br />

lately?) AI is not a new technology; researchers have been<br />

exploring AI since <strong>the</strong> early days 20 <strong>of</strong> computing. But,<br />

thanks to <strong>the</strong> exponential growth <strong>of</strong> <strong>the</strong> chip industry, as predicted<br />

so presciently by Gordon Moore in 1965, AI has come on <strong>the</strong><br />

scene now only because computers are finally 21 powerful enough –<br />

at everyday prices – to actually run <strong>the</strong> computation necessary to<br />

implement “AI.”<br />

As we have argued previously 22 :<br />

• “Machine learning techniques underlying personalised<br />

learning will disrupt, in <strong>the</strong> Christensen-sense, your<br />

child’s school in <strong>the</strong> near term. You can take that<br />

prediction to <strong>the</strong> bank!”<br />

For example, in early math education, in grades 1-3, companies<br />

such as Dreambox 23 <strong>and</strong> Knewton 24 are providing personalised<br />

instruction 25 systems that do result in improved test scores. That<br />

said, we urge <strong>the</strong> interested reader to review a “confession 26 ”<br />

by Larry Berger, CEO <strong>of</strong> Amplify 27 , Inc., who has been trying<br />

to build personalised learning systems for over 10 years. Berger<br />

argues that <strong>the</strong>re are a number <strong>of</strong> issues that prevent <strong>the</strong><br />

development <strong>of</strong> a personalised learning system for all grades <strong>and</strong><br />

all subjects <strong>and</strong> for all students.<br />

But, let’s not throw <strong>the</strong> baby out with <strong>the</strong> bath water: machine<br />

Machine learning techniques underlying<br />

personalised learning will disrupt, in <strong>the</strong><br />

Christensen-sense, your child’s school in<br />

<strong>the</strong> near term. You can take that prediction<br />

to <strong>the</strong> bank!<br />

learning algorithms can provide useful learning analytics to<br />

teachers so <strong>the</strong>y can personalize <strong>the</strong> learning for <strong>the</strong>ir students.<br />

While learning analytics can be deceiving at times,<br />

• ”… it has become clear to many that <strong>the</strong> use <strong>of</strong> data is critical<br />

to how ‘educators 28 evaluate <strong>the</strong>ir practices <strong>and</strong> monitor students’<br />

academic progress.’”<br />

As well, ML can provide more “natural” ways to interact with<br />

computers. For example, <strong>the</strong> explosion <strong>of</strong> voice-based interfaces<br />

(e.g., Apple’s Siri, Amazon’s Alexa 29 ) point to a future where<br />

voice can be used for both fast <strong>and</strong> sustained interaction. Integral<br />

to voice-based interfaces is natural language underst<strong>and</strong>ing 30<br />

by computer. Yes <strong>and</strong> no. As an experiment, we went into a<br />

kindergarten (children <strong>of</strong> age 4, 5, 6) <strong>and</strong> tried to use voice <strong>and</strong><br />

natural language s<strong>of</strong>tware to analyse <strong>the</strong>se children’s utterances.<br />

While <strong>the</strong> s<strong>of</strong>tware failed almost every time – kindergartners do<br />

not speak in grammatical sentences, do not enunciate clearly, <strong>and</strong><br />

make up words willy-nilly – we were assured by ML engineers that<br />

if <strong>the</strong>y had 5,000 hours <strong>of</strong> recorded speech from <strong>the</strong>se youngsters,<br />

<strong>the</strong>y could build an effective voice-based, natural language system<br />

for those youngsters! Perhaps!<br />

ML can also support <strong>the</strong> production <strong>of</strong> curriculum that serves as<br />

<strong>the</strong> heartbeat <strong>of</strong> a classroom.<br />

• Chi.Buzz – Developed at <strong>the</strong> UNESCO Mahatma G<strong>and</strong>hi<br />

Institute <strong>of</strong> <strong>Education</strong> for Peace <strong>and</strong> Sustainable Development<br />

(MGIEP), CHI – Collective Human <strong>Intelligence</strong> – is a platform<br />

for creating digital courses that employs ML to support <strong>the</strong><br />

development <strong>and</strong> enactment <strong>of</strong> digital courses.<br />

• Collabrify Roadmap – Developed at <strong>the</strong> Intergalactic Mobile<br />

Learning Center (IMLC), at Roadmap.center educators can create<br />

digital curriculum, Roadmaps, visual, concept-map-like structures,<br />

that guide students through a lesson. Finding content on <strong>the</strong><br />

Internet (in <strong>the</strong> form <strong>of</strong> OER – Open <strong>Education</strong> Resources) for<br />

<strong>the</strong> learning activities in those Roadmaps is a challenging, timeconsuming<br />

activity. We are exploring how to incorporate ML to<br />

provide curriculum developers with suggestions about relevant<br />

OER content.<br />

Games, that combine novel, graphical interfaces with ML<br />

algorithms in <strong>the</strong> context <strong>of</strong> provocative story-lines, may ultimately<br />

lead <strong>the</strong> way in <strong>the</strong> digital transformation <strong>of</strong> K-12. Presenting<br />

learning opportunities in <strong>the</strong> context <strong>of</strong> solo <strong>and</strong> collaborative<br />

game play, may well better align with <strong>the</strong> motivating factors for<br />

today’s learners. As <strong>the</strong> learners <strong>of</strong> today play games outside <strong>of</strong><br />

school, it makes good sense to bring <strong>the</strong> mode <strong>of</strong> learning inside<br />

<strong>of</strong> school, as well.<br />

We truly believe that a bet against technological change is a really<br />

bad bet; new digital technologies will be invented <strong>and</strong> <strong>the</strong>y will<br />

enable dramatic transformations. But, while ML may have finally<br />

conquered 31 <strong>the</strong> game <strong>of</strong> Go <strong>and</strong> bested our best human player 32 ,<br />

delivering personalised instruction 33 , diagnosing disease, predicting<br />

financial – <strong>and</strong> wea<strong>the</strong>r – changes, etc. are highly challenging<br />

<strong>and</strong> will require our absolutely best thinking <strong>and</strong> absolutely best<br />

technology. <strong>The</strong>se are early days for ML; we urge patience <strong>and</strong><br />

caution 34 .<br />

Now, digital transformation isn’t just about technology! In all <strong>the</strong><br />

articles about DX, success in DX involves people, organizational<br />

culture, <strong>and</strong> leadership:<br />

• “Hire 35 <strong>the</strong> right people to bring about change. Create a culture<br />

that makes data-based decisions.”<br />

• “… success 36 requires CEOs to develop <strong>the</strong> right leadership<br />

capabilities, workforce skills <strong>and</strong> corporate cultures to support<br />

digital transformation….”<br />

• Easy to say! In today’s schools, teachers are predominately<br />

drawn from <strong>the</strong> “baby-boomer” <strong>and</strong> “Gen X” generations, who<br />

are digital immigrants, while <strong>the</strong> students in <strong>the</strong>ir classes are<br />

millennials who are digital natives.<br />

And 37 ,<br />

• “ … research shows that student teachers will initially teach <strong>the</strong><br />

way <strong>the</strong>y were taught.”<br />

While teachers can learn <strong>and</strong> change 38 , a recent study 39 found that:<br />

• “A majority (70-78%) <strong>of</strong> teachers expressed discomfort with<br />

newer teaching practices that rely on technology”<br />

Daunting! With all <strong>the</strong> pr<strong>of</strong>essional development that has<br />

been showered upon today’s teachers, “a majority” are still<br />

uncomfortable using technology to facilitate collaboration, enable<br />

project-based learning, personalize learning, etc.<br />

Concluding Remarks<br />

Digital transformation is not a one-time event; ra<strong>the</strong>r it is as<br />

ongoing process:<br />

• “Digital 40 is a journey that requires [organisations] to be agile”<br />

And,<br />

• “Digitally 41 maturing [organisations] cultivate a culture that<br />

supports change <strong>and</strong> attracts talent. Commitment <strong>and</strong> leadership<br />

are required. To succeed, you have to be dedicated to investment,<br />

change, talent development, <strong>and</strong> have a clear vision… <strong>and</strong> be<br />

committed to it.”<br />

In <strong>the</strong> business world:<br />

• “… all companies 42 must transform to be digital companies to<br />

remain competitive.”<br />

In K-12? <strong>The</strong> economic forces driving businesses are not<br />

necessarily <strong>the</strong> same ones that drive K-12 education. What<br />

schools, what districts will invest <strong>the</strong> time <strong>and</strong> resources to digitally<br />

transform? Interesting questions – stay tuned!<br />


ISSUE • 9<br />

1 1





1<br />

https://www.cisco.com/c/dam/en_us/solutions/industries/docs/education/digital-learningenvironment.pdf<br />

2<br />

https://www.avaya.com/en/faqs/faq-detail-digital-transformation-all-about/<br />

3<br />

http://dialog.vonage.com/industry-trends/<strong>the</strong>-5-key-drivers-<strong>of</strong>-digital- transformation-today<br />

4<br />

https://www.techrepublic.com/article/digital-transformation-technology-<strong>and</strong>- servicesspending-is-on-<strong>the</strong>-rise/<br />

5<br />

https://searchcio.techtarget.com/definition/SMAC-social-mobile-analytics-<strong>and</strong>-cloud<br />

6<br />

https://www.macrumors.com/2018/06/27/apple-streaming-service-bundle-tv-music-news/<br />

7<br />

https://www.avaya.com/en/faqs/faq-detail-digital-transformation-all-about/<br />

8<br />

https://www.npr.org/sections/ed/2018/04/29/604986823/what-a-nation-at-risk-got-wrong<strong>and</strong>-right-about-u-s-schools<br />

9<br />

https://www.npr.org/sections/ed/2018/04/29/604986823/what-a-nation-at-risk-got-wrong<strong>and</strong>-right-about-u-s-schools<br />

10<br />

http://www.govtech.com/education/k-12/2018-digital-school-district-survey-respondentsuse-technology-to-put-students-first.html<br />

11<br />

http://www.govtech.com/education/k-12/2018-digital-school-district-survey-respondentsuse-technology-to-put-students-first.html<br />

12<br />

https://<strong>the</strong>journal.com/articles/2018/08/13/personalized-learning-top-priority-for-techusage-in-k12.aspx<br />

13<br />

https://<strong>the</strong>journal.com/articles/2018/08/13/personalized-learning-top-priority-for-techusage-in-k12.aspx<br />

14<br />

https://www.cnn.com/2017/12/11/health/cell-phones-for-kids-parenting-without-bordersexplainer-intl/index.html<br />

15<br />

https://techcrunch.com/2016/05/19/<strong>the</strong>-average-age-for-a-child-getting-<strong>the</strong>ir-firstsmartphone-is-now-10-3-years/<br />

16<br />

https://<strong>the</strong>journal.com/articles/2015/09/21/oecd-report.aspx<br />

17<br />

http://dialog.vonage.com/industry-trends/<strong>the</strong>-5-key-drivers-<strong>of</strong>-digital-transformation-today<br />

18<br />

http://<strong>the</strong>learningcounsel.com/article/ai-vs-machine-learning-<strong>and</strong><br />

education%E2%80%99s-future<br />

19<br />

https://medium.com/datadriveninvestor/differences-between-ai-<strong>and</strong>-machine-learning-<strong>and</strong>why-it-matters-1255b182fc6<br />

20<br />

http://sitn.hms.harvard.edu/flash/2017/history-artificial-intelligence/<br />

21<br />

https://<strong>the</strong>journal.com/articles/2017/09/18/ai-is-on-fire.aspx<br />

22<br />

https://<strong>the</strong>journal.com/Articles/2017/07/24/A-Christensen-style-Disruption.aspx?Page=1<br />

23<br />

http://www.dreambox.com/white-papers/rti-toolkit?ppc&&gclid=Cj0KCQiA3b3gB<br />

RDAARIsAL6D-N-xZaXNmroovmtNBtu6x_ZRazIbLkJ2qaySOzEt8p1wEXIT-tEq-<br />

BkaAkR7EALw_wcB&gclsrc=aw.ds<br />

24<br />

https://www.knewton.com/<strong>the</strong>-power-<strong>of</strong>-altas-adaptive-technology/<br />

25<br />

https://<strong>the</strong>journal.com/articles/2014/12/04/personalized-learning.aspx<br />

26<br />

http://blogs.edweek.org/edweek/rick_hess_straight_up/2018/02/a_confession_<strong>and</strong>_a_<br />

question_on_personalized_learning.html<br />

27<br />

https://www.amplify.com/<br />

28<br />

https://ies.ed.gov/ncee/wwc/Docs/PracticeGuide/dddm_pg_092909.pdf<br />

29<br />

https://developer.amazon.com/alexa-skills-kit/vui<br />

30<br />

https://www.nytimes.com/2018/11/18/technology/artificial-intelligence-language.html<br />

31<br />

https://spectrum.ieee.org/semiconductors/processors/<strong>the</strong>-multiple-lives-<strong>of</strong>-moores-law<br />

32<br />

https://techcrunch.com/2017/05/23/googles-alphago-ai-beats-<strong>the</strong>-worlds-best-human-goplayer/<br />

33<br />

https://nepc.colorado.edu/publication/newsletter-personalized-learning-121118<br />

34<br />

http://www.rogerschank.com/fraudulent-claims-made-by-IBM-about-Watson-<strong>and</strong>-AI<br />

35<br />

https://hbr.org/2016/03/<strong>the</strong>-industries-that-are-being-disrupted-<strong>the</strong>-most-by-digital<br />

36<br />

https://www.weforum.org/agenda/2016/01/digital-disruption-has-only-just-begun/<br />

37<br />

https://web.uvic.ca/~thopper/Cupr/Archived/waytaught.htm<br />

38<br />

https://gianfrancoconti.wordpress.com/2015/12/23/why-teachers-teach-<strong>the</strong>-way-<strong>the</strong>y-dopart-1/<br />

39<br />

https://<strong>the</strong>journal.com/articles/2018/07/30/educators-largely-uncomfortable-with-newertech-based-teaching-practices.aspx<br />

40<br />

https://www.zdnet.com/article/digital-transformation-retooling-business-for-a-new-age/<br />

41<br />

https://sloanreview.mit.edu/projects/achieving-digital-maturity/<br />

42<br />

http://dialog.vonage.com/cloud/5-myths-about-digital-transformation<br />

ISSUE • 9<br />

1 3

opinion<br />

Peace Machine as a<br />


Re-emerging G<strong>and</strong>hi<br />


A<br />

rtificial <strong>Intelligence</strong> (AI) has been under<br />

development for more than 50 years.<br />

Many ideas central to AI have been under<br />

consideration from very early on. <strong>The</strong> basic<br />

idea is to model or mimic human intelligence, including decision<br />

making, problem solving, language <strong>and</strong> speech underst<strong>and</strong>ing,<br />

machine learning, pattern recognition <strong>and</strong> a number <strong>of</strong> o<strong>the</strong>r<br />

characteristics. Even emotions <strong>and</strong> creativity have been taken<br />

into consideration. A common idea among AI researchers is to<br />

see <strong>the</strong> developments as models <strong>of</strong> <strong>the</strong> true phenomenon, not as<br />

independent achievements. Computers are not conscious <strong>and</strong><br />

emotional in a human way but successful analysis <strong>and</strong> modelling<br />

has causal effects in <strong>the</strong> world <strong>of</strong> living humans.<br />

Timo Honkela serves as a Pr<strong>of</strong>essor at <strong>the</strong> University<br />

<strong>of</strong> Helsinki, Department <strong>of</strong> Digital Humanities.<br />

Earlier positions include Pr<strong>of</strong>essorships at <strong>the</strong><br />

University <strong>of</strong> Art <strong>and</strong> Design Helsinki <strong>and</strong> Helsinki<br />

University <strong>of</strong> Technology, where he received his PhD<br />

in 1998. Dr. Honkela has served as Chair <strong>of</strong> Finnish<br />

<strong>Artificial</strong> <strong>Intelligence</strong> Society <strong>and</strong> as Executive<br />

Board Member <strong>of</strong> <strong>the</strong> European Neural Networks<br />

Society. His interests are in philosophical topics<br />

regarding human existence, emergence in relation<br />

to conceptual systems, <strong>and</strong> cultural evolution.<br />

<strong>The</strong> societal impact <strong>of</strong> AI has grown considerably during<br />

recent years. <strong>The</strong>re are multiple reasons for this development.<br />

<strong>The</strong> use <strong>of</strong> artificial neural networks <strong>and</strong> statistical machine<br />

learning are used to analyse large data collections. <strong>The</strong>se collections<br />

are like artificial experiences: sounds, spoken <strong>and</strong> written language,<br />

images <strong>and</strong> videos. <strong>The</strong> systems have growing cognitive capacity,<br />

ability to model human intellect based on <strong>the</strong> availability <strong>of</strong><br />

increased amount <strong>of</strong> data <strong>and</strong> computing power <strong>and</strong> data storage.<br />

Earlier <strong>the</strong> systems were programmed with <strong>the</strong> aim to transfer<br />

expertise into <strong>the</strong>m in <strong>the</strong> form <strong>of</strong> rules. Current solutions are<br />

mainly based on programming systems to learn.<br />

<strong>The</strong> AI developments are mostly directed to industrial <strong>and</strong><br />

commercial purposes. <strong>The</strong> steps taken in this may lead into more<br />

efficient provision <strong>of</strong> products <strong>and</strong> services. An emerging challenge<br />

is that many pr<strong>of</strong>essionals lose <strong>the</strong>ir work. This leads into problems<br />

in <strong>the</strong> societies as finding <strong>the</strong> work place as a replacement is not<br />

straightforward. <strong>The</strong> AI methodologies are also used to develop<br />

increasingly efficient weapons. Overall, many people feel that <strong>the</strong><br />

<strong>The</strong> Peace Machine is a concept <strong>of</strong><br />

applying AI with a range <strong>of</strong> technologies<br />

<strong>and</strong> disciplines underst<strong>and</strong>ing to promote<br />

peaceful conditions in <strong>the</strong> world.<br />

world is becoming increasingly frightening,<br />

also due to growing environmental<br />

concerns.<br />

<strong>The</strong> Peace Machine is a concept<br />

<strong>of</strong> applying AI with a range <strong>of</strong><br />

technologies <strong>and</strong> disciplines<br />

underst<strong>and</strong>ing to promote peaceful<br />

conditions in <strong>the</strong> world. <strong>The</strong><br />

main areas are (1) increasing mutual<br />

underst<strong>and</strong>ing among people, (2)<br />

promoting awareness <strong>of</strong> one’s own <strong>and</strong><br />

o<strong>the</strong>rs’ emotions, <strong>and</strong> (3) developing fair<br />

conditions on a large scale. <strong>The</strong> Peace<br />

Machine is not one device or system<br />

but a wide range <strong>of</strong> modules, tools<br />

<strong>and</strong> applications. <strong>The</strong>y are mostly<br />

intended to be available for <strong>and</strong> used by<br />

large numbers <strong>of</strong> people.<br />

Development <strong>of</strong> <strong>the</strong> concept <strong>of</strong> <strong>the</strong><br />

Peace Machine began in early 2017. <strong>The</strong><br />

author’s own personal history adds ano<strong>the</strong>r<br />

dimension to this. This includes 30 years<br />

<strong>of</strong> research on AI <strong>and</strong> being diagnosed<br />

with <strong>the</strong> worst kind <strong>of</strong> brain cancer. <strong>The</strong><br />

tumor was operated upon in December<br />

2014 but left <strong>the</strong> author blind in <strong>the</strong> right<br />

eye. Never<strong>the</strong>less, <strong>the</strong> fact that life gave <strong>the</strong><br />

author a reprieve led to <strong>the</strong> idea <strong>of</strong> a book<br />

on <strong>the</strong> Peace Machine. Many interviews<br />

to newspapers, radio <strong>and</strong> television<br />

resulted in an outpouring <strong>of</strong> support, <strong>and</strong><br />

thous<strong>and</strong>s <strong>of</strong> people helped <strong>the</strong> author<br />

write <strong>the</strong> book “<strong>The</strong> Peace Machine -<br />

<strong>The</strong> Testament <strong>of</strong> an AI Researcher”,<br />

which was published in October 2017 in<br />

<strong>the</strong> Finnish language. <strong>The</strong> book is now<br />

being translated to English <strong>and</strong> should be<br />

published by 2019. Ano<strong>the</strong>r effect <strong>of</strong> this<br />

was that a society in Helsinki came into<br />

being in September 2018 to coordinate<br />

research, development, <strong>and</strong> international<br />

collaboration on this topic.<br />

ISSUE • 9<br />

1 5


Increasing mutual underst<strong>and</strong>ing among<br />

people<br />

In general, it is hard to arrive at mutual underst<strong>and</strong>ing.<br />

It is particularly difficult to achieve this when people have<br />

different backgrounds, education, experience, cultural<br />

contexts, etc. People not only have different values <strong>and</strong><br />

opinions, <strong>the</strong> same words also hold different meanings<br />

for <strong>the</strong>m.This is quite obvious for those who work in multicultural<br />

or international organisations. Information technology has not<br />

been aligned to deal with this phenomenon as computers are<br />

programmed with clear definitions <strong>of</strong> meaning. <strong>The</strong> development<br />

<strong>of</strong> AI <strong>and</strong> information technology has to be guided by careful,<br />

philosophical underst<strong>and</strong>ing. <strong>The</strong> systems that conduct meaning<br />

negotiations take into account <strong>the</strong> contexts in which people express<br />

Luckily, in addition to machine<br />

translation, tools for meaning negotiation<br />

can be developed to help in managing<br />

situations in which differences in<br />

experience, education <strong>and</strong> values have<br />

caused <strong>the</strong> conceptual systems <strong>of</strong> people to<br />

be different.<br />

<strong>the</strong>mselves. <strong>The</strong> meaning <strong>of</strong> a word is not carved in stone but it is<br />

dependent on <strong>the</strong> linguistic, external <strong>and</strong> even internal context. Us<br />

humans can deal with this to some extent but information systems<br />

usually cannot deal with this at all. <strong>The</strong> idea <strong>of</strong> unambiguous<br />

meaning <strong>and</strong> crisp interpretation has been a guiding principle for<br />

many str<strong>and</strong>s <strong>of</strong> philosophy <strong>and</strong> for computer science. Regrettably,<br />

it is not practical because we humans form heterogeneous<br />

groups.. Luckily, in addition to machine translation, tools<br />

for meaning negotiation can be developed to help in<br />

managing situations in which differences in experience,<br />

education <strong>and</strong> values have caused <strong>the</strong> conceptual<br />

systems <strong>of</strong> people to be different. <strong>The</strong> ways in which people<br />

underst<strong>and</strong> language cannot be solved by education because us<br />

human cannot be reprogrammed like machines. <strong>The</strong> machines<br />

can, however, serve as our interpreters if we so wish.<br />

Good quality machine translation hypo<strong>the</strong>tically improves<br />

peaceful conditions in <strong>the</strong> world as people can encounter over<br />

language borders. On <strong>the</strong> o<strong>the</strong>r h<strong>and</strong>, meaning negotiations may<br />

be needed in addition to straightforward machine translation.<br />

Google has succeeded in improving <strong>the</strong> quality <strong>of</strong> its machine<br />

translation system by taking into use neural network methodology,<br />

in particular deep learning systems with multiple “intuitive”<br />

information processing layers. <strong>The</strong> Peace Machine is intended<br />

to be available <strong>and</strong> useful for people all around <strong>the</strong> world. To<br />

reach necessary levels <strong>of</strong> trust, <strong>the</strong> systems must use open source<br />

s<strong>of</strong>tware. A system that can have substantial effect on people<br />

cannot be closed within <strong>the</strong> digital walls <strong>of</strong> <strong>the</strong> system. <strong>The</strong><br />

citizens <strong>and</strong> <strong>the</strong> governments, or <strong>the</strong>ir trusted representatives need<br />

to be able to see what is within <strong>the</strong> system, whe<strong>the</strong>r <strong>the</strong>y live in<br />

India, USA, China, Russia, Iran, Israel, Icel<strong>and</strong> or anywhere else.<br />

Promoting awareness <strong>of</strong> one’s own <strong>and</strong><br />

o<strong>the</strong>rs’ emotions<br />

An important part <strong>of</strong> human decision-making is based<br />

on our emotions. Emotions help us to make quick decisions<br />

when swift action is required. On <strong>the</strong> o<strong>the</strong>r h<strong>and</strong>, our emotions can<br />

become tuned into some less optimal states or reaction tendencies.<br />

From <strong>the</strong> point <strong>of</strong> view <strong>of</strong> promoting peaceful conditions,<br />

orientation towards fear <strong>and</strong> anger is problematic. <strong>The</strong>y prepare<br />

us to be aggressive <strong>and</strong> violent. <strong>The</strong>y can be used to manipulate<br />

us to be available for violent actions even when <strong>the</strong>se actions are<br />

against our best interests. It can be also asked if violence is ever<br />

truly well grounded. <strong>The</strong> true problems in <strong>the</strong> world are related<br />

to poverty, health <strong>and</strong> <strong>the</strong> environment. Violence <strong>and</strong> war may<br />

sometimes be seen as a means to break through some situation but<br />

<strong>the</strong>re are o<strong>the</strong>r solutions. Machines do not have emotions<br />

but <strong>the</strong>y can analyse our emotional states <strong>and</strong> related<br />

individual <strong>and</strong> social processes. We can develop our wisdom<br />

with <strong>the</strong> AI systems that would make us more aware <strong>of</strong> underlying<br />

mental processes that may be problematic, for instance, due to<br />

childhood trauma or culturally embedded reaction patterns based<br />

on cultural patterns established over generations. It may sound<br />

paradoxical but we could use computers to help us to achieve<br />

progress in <strong>the</strong>se matters. A high level <strong>of</strong> transparency is needed<br />

Machines do not have emotions but <strong>the</strong>y<br />

can analyse our emotional states <strong>and</strong><br />

related individual <strong>and</strong> social processes.<br />

in <strong>the</strong> development <strong>of</strong> <strong>the</strong>se kinds <strong>of</strong> systems. Machine learning<br />

is not a straightforward solution in all <strong>of</strong> this because machine<br />

learning, in its most common forms, is based on data that may<br />

also lead to problematic solutions, not only good ones. <strong>The</strong>refore,<br />

intentionality is an essential concept: what do we wish to achieve,<br />

why <strong>and</strong> how. This is only to a limited way a technological<br />

question but much more a societal <strong>and</strong> ethical question.<br />

AI can be used <strong>and</strong> fur<strong>the</strong>r developed into<br />

directions that support reaching goals that<br />

Mahatma G<strong>and</strong>hi would be happy <strong>and</strong><br />

proud <strong>of</strong>.<br />

Developing fair conditions on a large scale<br />

What is fair can be seen as a question <strong>of</strong> meaning negotiation.<br />

Moreover, it is an essential question when we consider human<br />

condition <strong>and</strong> our societies. Situations in different parts <strong>of</strong> <strong>the</strong> world<br />

are history dependant to a high degree. Not all people consider<br />

all people to be free <strong>and</strong> have basic human rights regardless <strong>of</strong><br />

<strong>the</strong>ir gender, race <strong>and</strong> position. From <strong>the</strong> systems <strong>the</strong>oretical<br />

point <strong>of</strong> view, equality is a problematic notion. One possibility<br />

<strong>of</strong> approaching this is to consider distributions <strong>and</strong> contextual<br />

representations <strong>of</strong> reality. Like emerging new medicine provides a<br />

chance to take care <strong>of</strong> each person individually, <strong>the</strong> societies can<br />

be made better for all taking each person individually, providing<br />

improved conditions for all. This can help in <strong>the</strong> promotion <strong>of</strong><br />

freedom, associated with increasing level <strong>of</strong> responsibility. <strong>The</strong><br />

developments do not need to be directed by a small number<br />

<strong>of</strong> people like it is done nowadays. Namely, natural language<br />

processing <strong>and</strong> machine learning is making it possible to organise<br />

true meetings among very large number <strong>of</strong> people. A meeting <strong>of</strong> a<br />

million or billion people is not a voting arena but provides chances<br />

for true discussions when matters are considered among people<br />

at a grass root level. People will be able to craft solutions with<br />

unforeseen capacity <strong>and</strong> level <strong>of</strong> participation.<br />

Some <strong>of</strong> <strong>the</strong> goals for development may sound utopian. However,<br />

<strong>the</strong> content presented in <strong>the</strong> book does not relate to predictions but<br />

to goals <strong>and</strong> plans. AI can be used <strong>and</strong> fur<strong>the</strong>r developed<br />

into directions that support reaching goals that<br />

Mahatma G<strong>and</strong>hi would be happy <strong>and</strong> proud <strong>of</strong>.<br />


ISSUE • 9<br />

1 7

opinion<br />

Promoting Peace: How Can<br />


Data Science Assist?<br />


Srujana is a Machine Learning researcher <strong>and</strong><br />

consultant with over a decade <strong>of</strong> experience. She<br />

currently works as a founding member <strong>of</strong> a stealth<br />

mode startup. Prior to that, she was a principal data<br />

scientist at Flipkart (Bangalore) <strong>and</strong> a volunteer<br />

for Ekstep, an education startup. She has been<br />

employed with Machine Learning groups at Amazon<br />

(Bangalore), IBM Research (Bangalore/New Delhi),<br />

<strong>and</strong> Yahoo Research (Santa Clara). Srujana has<br />

published her work in several conferences <strong>and</strong><br />

journals on data mining/machine learning <strong>and</strong> is<br />

<strong>the</strong> recipient <strong>of</strong> multiple best paper awards. She<br />

received her M.S. <strong>and</strong> Ph.D. from <strong>the</strong> University<br />

<strong>of</strong> Texas at Austin <strong>and</strong> her B. Tech. degree from IIT<br />

Madras.<br />

P<br />

eace has long been a pr<strong>of</strong>essed goal <strong>of</strong> <strong>the</strong><br />

mankind, but it has not been pursued with<br />

<strong>the</strong> rigour it dem<strong>and</strong>s. While many intergovernmental<br />

organisations <strong>and</strong> policy institutes<br />

do exist, which are dedicated to promoting peace, <strong>the</strong> study <strong>of</strong><br />

peace-building <strong>and</strong> conflict resolution mechanisms has primarily<br />

been limited to <strong>the</strong> humanities community with only peripheral<br />

involvement from experts in data science <strong>and</strong> related areas.<br />

Clearly, <strong>the</strong> necessity to account for cultural nuances, human<br />

emotions, irrational behavior, <strong>and</strong> different notions <strong>of</strong> fairness<br />

in peace-building makes it a highly challenging space for a hard<br />

scientific approach. However, <strong>the</strong> effectiveness <strong>of</strong> data science<br />

approaches has been demonstrated in domains with similar<br />

complexity such as advertising, online political campaigning,<br />

personalised learning, <strong>and</strong> various forms <strong>of</strong> match-making. <strong>The</strong><br />

recent Facebook-Cambridge Analytica sc<strong>and</strong>al, in particular,<br />

highlights <strong>the</strong> potency <strong>of</strong> a data-driven approach to influence<br />

human behavior at a massive scale. <strong>The</strong> primary reason for<br />

<strong>the</strong> limited participation <strong>of</strong> <strong>the</strong> data science community<br />

in <strong>the</strong> pursuit <strong>of</strong> peace is that it has not has been<br />

sufficiently incentivised in our increasingly capitalistic<br />

world unlike advertising, which is now being honed to almost an<br />

exact science at companies such as Google <strong>and</strong> Facebook.<br />

This article attempts to examine certain vital elements required<br />

to advance peace <strong>and</strong> provides a sense <strong>of</strong> how data science can<br />

potentially contribute to those elements. At <strong>the</strong> same time, it lists<br />

caveats <strong>and</strong> external factors that need to be considered to ensure<br />

that <strong>the</strong> potential is realised.<br />

Elements <strong>of</strong> peaceful conflict resolution<br />

Heading home one evening in a cab, I had a close encounter<br />

with mob violence. A car crashed into our vehicle, <strong>and</strong> within a<br />

matter <strong>of</strong> minutes, a massive crowd <strong>of</strong> people assembled. Folks<br />

started arguing with each o<strong>the</strong>r, <strong>and</strong> before long, <strong>the</strong> arguments<br />

devolved into a violent fight resulting in multiple people getting<br />

injured without any resolution on h<strong>and</strong>ling <strong>the</strong> damages from<br />

<strong>the</strong> accident. This incident took me back to a more serious<br />

car accident that happened a few years ago in California. In<br />

contrast to <strong>the</strong> episode in Bangalore, <strong>the</strong> parties involved in<br />

<strong>the</strong> accident just pulled over, assessed <strong>the</strong> damages, exchanged<br />

insurance numbers, <strong>and</strong> called <strong>the</strong> authorities to lodge a FIR. We<br />

commiserated with each o<strong>the</strong>r on <strong>the</strong> unfortunate incident <strong>and</strong><br />

wished each o<strong>the</strong>r safe journey on our way back home. Within two<br />

days, <strong>the</strong> insurance companies had figured out how <strong>the</strong> damages<br />

had to be paid for, <strong>and</strong> both parties readily accepted <strong>the</strong> decision.<br />

<strong>The</strong> contrast between <strong>the</strong> two incidents could not have been<br />

more stark. <strong>The</strong> triggers were <strong>the</strong> same, but <strong>the</strong> outcomes vastly<br />

different, which made me wonder why that was <strong>the</strong> case.<br />

<strong>The</strong>re seemed to be at least four salient factors that could explain<br />

<strong>the</strong>se differences.<br />

• Regulatory structures that strongly discourage violence<br />

• Predisposition <strong>of</strong> <strong>the</strong> population towards peaceful behaviour<br />

• Familiarity with <strong>the</strong> resolution process <strong>of</strong> common conflicts<br />

• Availability <strong>of</strong> arbiters perceived as neutral <strong>and</strong> knowledgeable<br />

by all <strong>the</strong> parties<br />

None <strong>of</strong> <strong>the</strong> above positive factors are easy to create. Nei<strong>the</strong>r is<br />

this list exhaustive as <strong>the</strong>re are many o<strong>the</strong>r aspects (e.g., inherent<br />

power imbalances, malicious intent by third parties) that one needs<br />

to consider to address <strong>the</strong> vast variety <strong>of</strong> conflicts such as domestic<br />

violence, communal riots <strong>and</strong> territorial disputes. However, <strong>the</strong><br />

good news is that over <strong>the</strong> last two decades, <strong>the</strong>re have been<br />

massive advances in data management, internet technologies,<br />

computing technologies, <strong>and</strong> machine learning, <strong>and</strong> <strong>the</strong>se<br />

technologies have leveraged each o<strong>the</strong>r to create a perfect storm,<br />

which makes it possible for us to attempt to solve such challenging<br />

problems.<br />

How can data science assist?<br />

To underst<strong>and</strong> how <strong>and</strong> where data science <strong>and</strong> machine learning<br />

(ML) can help, let us delve deeper into <strong>the</strong> obstacles faced in<br />

creating some <strong>of</strong> <strong>the</strong> factors that contribute to peace.<br />

Regulatory structures & policy decisions. Government<br />

organisations <strong>and</strong> policy think tanks have enormous<br />

power to influence policy decisions <strong>and</strong> regulatory<br />

mechanisms, but <strong>of</strong>ten do not make good choices even with<br />

an abundance <strong>of</strong> good intentions. <strong>The</strong>se suboptimal choices are<br />

primarily due to a lack <strong>of</strong> access to <strong>the</strong> relevant data as well as an<br />

inability to distill <strong>the</strong> available data to perform a what-if-analysis,<br />

e.g., is free public schooling till <strong>the</strong> age <strong>of</strong> 18 years likely to result<br />

ISSUE • 9<br />

1 9

opinion<br />

in fewer cases <strong>of</strong> domestic violence?<br />

Data science approaches can<br />

optimise this decision-making<br />

process in multiple ways by:<br />

• enabling semi-automated consolidation<br />

<strong>of</strong> relevant data from a wide range<br />

<strong>of</strong> heterogeneous, structured <strong>and</strong><br />

unstructured multi-lingual, multi-media<br />

sources across <strong>the</strong> world,<br />

• <strong>of</strong>fering insights on temporal trends,<br />

natural clusters, <strong>and</strong> frequent patterns to<br />

policymakers,<br />

• discovering associations between policy<br />

choices <strong>and</strong> outcomes for different<br />

contexts via analytic models learned<br />

from <strong>the</strong> data,<br />

• performing quick <strong>and</strong> accurate<br />

evaluation <strong>of</strong> different policy choices via<br />

statistical tests.<br />

Promoting a peaceful mindset<br />

among people. Today’s cultural<br />

forces (e.g., social media, news<br />

media, video games, movies)<br />

<strong>and</strong> education systems are much<br />

more geared towards promoting<br />

competition, aggression, <strong>and</strong><br />

violence than peacefulness <strong>and</strong><br />

cooperation. <strong>The</strong> current state <strong>of</strong> affairs<br />

most likely is due to <strong>the</strong> apathy on <strong>the</strong><br />

part <strong>of</strong> <strong>the</strong> educators <strong>and</strong> media owners<br />

towards promoting peace. Never<strong>the</strong>less,<br />

even if a positive intent could be induced<br />

with changes in regulation, it is still<br />

non-trivial for moderators to weed out<br />

objectionable content <strong>and</strong> for educators<br />

<strong>and</strong> influencers to take appropriate<br />

actions to mold <strong>the</strong> behaviour <strong>of</strong> people<br />

under <strong>the</strong>ir influence since that requires<br />

fine-grained contextualisation.<br />

Data-science based systems can<br />

be quite helpful in solving some<br />

<strong>of</strong> <strong>the</strong>se challenges. In particular,<br />

it is possible to automatically assess<br />

different types <strong>of</strong> content (movies,<br />

games, online posts) on <strong>the</strong> likely impact<br />

it has on a consumer, which can <strong>the</strong>n<br />

be used to regulate <strong>the</strong> content sources.<br />

Twitter, Facebook, <strong>and</strong> Google already<br />

have such systems in place. Analysis <strong>of</strong><br />

behavioural data, combined with medical,<br />

Today’s cultural forces<br />

(e.g., social media, news<br />

media, video games,<br />

movies) <strong>and</strong> education<br />

systems are much<br />

more geared towards<br />

promoting competition,<br />

aggression, <strong>and</strong> violence<br />

than peacefulness <strong>and</strong><br />

cooperation.<br />

demographic, environmental factors <strong>and</strong><br />

content consumption history, can also<br />

pinpoint <strong>the</strong> reasons for a peaceful or<br />

violent disposition, which can <strong>the</strong>n be<br />

used to recommend appropriate corrective<br />

or learning actions to educators/leaders<br />

targeted towards each individual. For<br />

example, if a child is aggressive because<br />

<strong>of</strong> a lack <strong>of</strong> basic needs such as food<br />

<strong>and</strong> housing, it is preferable to focus on<br />

addressing those issues than inundating<br />

her/him with messages on peacefulness.<br />

Familiarity with common resolution<br />

mechanisms. A common cause for<br />

avoidable violence is ignorance <strong>of</strong><br />

resolution mechanisms for everyday<br />

conflicts, e.g., domestic abuse <strong>of</strong> women<br />

by <strong>the</strong>ir spouses, communal riots,<br />

economic exploitation <strong>of</strong> tribal <strong>and</strong><br />

rural people by o<strong>the</strong>rs. This observation<br />

is particularly true for <strong>the</strong> poor <strong>and</strong><br />

under-educated parts <strong>of</strong> <strong>the</strong> population.<br />

ML - based personalised content<br />

recommendation systems that can<br />

<strong>of</strong>fer content on conflict resolution<br />

processes tailored to a person’s<br />

situation, in <strong>the</strong> language <strong>and</strong><br />

media <strong>of</strong> his/her choice can go<br />

a long way in addressing this<br />

problem. One can also create engaging<br />

role-playing games powered by ML<br />

algorithms that can simulate typical<br />

conflict scenarios <strong>and</strong> train <strong>the</strong> players on<br />

navigating such situations in real world.<br />

Neutral, knowledgeable <strong>and</strong><br />

effective arbiters. Resolution <strong>of</strong><br />

conflicts is <strong>of</strong>ten tricky in <strong>the</strong> absence <strong>of</strong><br />

a neutral, knowledgeable arbiter. In most<br />

conflicts, it is not only <strong>the</strong> neutrality that is<br />

in question; <strong>the</strong> sheer volume <strong>of</strong> disputes<br />

makes even <strong>the</strong> number <strong>of</strong> available<br />

arbiters extremely low. Automated ML<br />

agents by design do not have a bias<br />

specific to a particular situation <strong>and</strong> are<br />

more likely to be neutral <strong>and</strong> perceived<br />

as such. It is also a scalable option to<br />

deploy automated agents directly or as<br />

assistants to humans to enable thoughtful<br />

<strong>and</strong> timely resolution <strong>of</strong> conflicts. <strong>The</strong>re<br />

is, in fact, ongoing work on using<br />

ML models to assist judges in<br />

coming up with objective <strong>and</strong> fair<br />

judgments. Specialised ML systems are<br />

also being used to evaluate <strong>the</strong> truth <strong>of</strong><br />

objective claims from different sources<br />

based on <strong>the</strong> raw evidence, which can<br />

help in resolving differences.<br />

<strong>The</strong> above ideas on applications <strong>of</strong> data<br />

science for peaceful conflict resolution are<br />

just <strong>the</strong> tip <strong>of</strong> <strong>the</strong> iceberg, <strong>and</strong> <strong>the</strong>re are<br />

possibly many more.<br />

Caveats on <strong>the</strong> limits <strong>of</strong><br />

data science<br />

Intelligent ML-based systems such as<br />

Google search, Facebook feeds, Amazon<br />

product recommendations <strong>and</strong> Uber<br />

have become an integral part <strong>of</strong> our lives,<br />

but most <strong>of</strong> us do not yet have a good<br />

underst<strong>and</strong>ing <strong>of</strong> how data science <strong>and</strong><br />

ML work. Misconceptions around data<br />

science make it seem like ei<strong>the</strong>r like a<br />

“panacea” or an “out-<strong>of</strong>-control beast”,<br />

leading to all kinds <strong>of</strong> absurd headlines<br />

full <strong>of</strong> hype or fear. <strong>The</strong> reality is far<br />

different, <strong>and</strong> it is essential to underst<strong>and</strong><br />

this to ensure appropriate <strong>and</strong> effective<br />

use <strong>of</strong> data science.<br />

<strong>The</strong> core <strong>of</strong> data science is machine<br />

learning, which comprises techniques<br />

that enable machines to learn from data<br />

without being explicitly programmed.<br />

<strong>The</strong>se techniques are different from <strong>the</strong><br />

older generation AI methods such as rulebased<br />

systems that had to be explicitly<br />

programmed by experts. In fact, ML methods have more in<br />

common with <strong>the</strong> way humans learn. For example, consider how<br />

a child learns. To him/her, it is important to determine if a new<br />

person is friendly or hostile. <strong>The</strong> learning process involves making<br />

multiple observations over time on people’s attributes such as <strong>the</strong><br />

tone <strong>of</strong> voice, facial expressions, <strong>and</strong> creating a mental model <strong>of</strong><br />

hostile vs. friendly, which can be <strong>the</strong>n be employed to characterize<br />

a new person.<br />

Machine learning operates similarly <strong>and</strong> involves three key<br />

elements.<br />

• Explicit representation <strong>of</strong> all <strong>the</strong> relevant information, e.g., <strong>the</strong><br />

tone <strong>of</strong> voice, as variables in a ma<strong>the</strong>matical framework<br />

• Collection <strong>of</strong> empirical observations on <strong>the</strong> variables<br />

• Learning algorithms to infer associations between variables using<br />

<strong>the</strong> ga<strong>the</strong>red data.<br />

<strong>The</strong> benefits <strong>of</strong> data science methods relative to manual judgment<br />

stem from <strong>the</strong>ir ability to effortlessly h<strong>and</strong>le trade-<strong>of</strong>fs among<br />

multiple possibly conflicting signals, incorporate massive amounts<br />

<strong>of</strong> data, <strong>and</strong> contextualise predictions for fine-grained contexts.<br />

However, one must be extremely mindful <strong>of</strong> <strong>the</strong> fact that ML<br />

techniques have a heavy dependence on data <strong>and</strong> do not have<br />

any innate ability to generate predictions. In particular, <strong>the</strong> most<br />

commonly used class <strong>of</strong> ML methods, i.e., supervised learning<br />

methods, need adequately labelled data to first learn what <strong>the</strong>y are<br />

expected to predict, e.g. if we want to rate content as peaceful or<br />

violent, we first need sufficient training examples <strong>of</strong> peaceful <strong>and</strong><br />

violent content. Fur<strong>the</strong>rmore, <strong>the</strong> applicability <strong>of</strong> an ML model<br />

is <strong>of</strong>ten limited to data distribution it is trained on, e.g., models<br />

trained on <strong>the</strong> US population might or might not generalize to<br />

a different community. If it turns out that data used to build ML<br />

models is biased or noisy, <strong>the</strong>n <strong>the</strong> predictions <strong>of</strong> <strong>the</strong> resulting<br />

models are also likely to be biased <strong>and</strong> erroneous.<br />

External regulatory factors needed for<br />

success<br />

Like any o<strong>the</strong>r powerful technology, data science is<br />

a sharp double-edged sword that can cut both ways.<br />

Consolidation <strong>of</strong> data not only unlocks valuable insights but<br />

also poses enormous risks if it falls into <strong>the</strong> h<strong>and</strong>s <strong>of</strong> malicious<br />

players. It is imperative, <strong>the</strong>refore, to carefully formulate policies<br />

that safeguard against privacy breaches <strong>and</strong> potential misuse, <strong>and</strong><br />

ensure strict compliance. <strong>The</strong> GDPR regulation is a commendable<br />

effort in this direction, but more work remains to be done.<br />

<strong>The</strong>re is considerable evidence to demonstrate that<br />

privately-owned information ecosystems such as search<br />

engines, social media networks <strong>and</strong> news media, driven<br />

by engagement-drive business metrics, are resulting in<br />

increasingly polarised communities. Data science methods<br />

are being aggressively deployed in <strong>the</strong>se ecosystems but primarily<br />

to optimise business metrics, which are not necessarily aligned<br />

towards creating a peaceful society. How much ever governmental<br />

<strong>and</strong> inter-governmental organisations attempt to promote peace,<br />

Like any o<strong>the</strong>r powerful technology, data<br />

science is a sharp double-edged sword that<br />

can cut both ways.<br />

<strong>the</strong>se efforts are likely to be subverted by <strong>the</strong>se highly influential<br />

ecosystems. It is critical to impose well-thought-out content<br />

st<strong>and</strong>ards <strong>and</strong> regulatory mechanisms that make it economically<br />

pr<strong>of</strong>itable for <strong>the</strong>se privately-owned ecosystems to promote peace.<br />

Summing up<br />

To conclude, data science has <strong>the</strong> potential to make immense<br />

contributions to <strong>the</strong> pursuit <strong>of</strong> peace. Providing macro-level<br />

insights to policymakers, curating personalised content on peaceful<br />

conflict resolution, guiding educators in molding <strong>the</strong> behaviour <strong>of</strong><br />

<strong>the</strong>ir wards <strong>and</strong> automated agents for arbitration are some <strong>of</strong> <strong>the</strong><br />

many ways that data science can assist in promoting peace. <strong>The</strong>re<br />

are, <strong>of</strong> course, caveats on <strong>the</strong> applicability <strong>of</strong> data science <strong>and</strong> risks<br />

<strong>of</strong> privacy breaches that should be taken into consideration. Most<br />

importantly, it is essential to harness <strong>the</strong> capitalistic<br />

forces <strong>of</strong> <strong>the</strong> world towards <strong>the</strong> pursuit <strong>of</strong> peace.<br />


ISSUE • 9<br />

2 1

opinion<br />

Declaration <strong>of</strong><br />

Independence II<br />



Data Subjects <strong>of</strong> all countries unite !<br />

We were promised ultimate freedom <strong>and</strong> liberty <strong>and</strong> instead we have<br />

become serfs <strong>of</strong> new masters.<br />

Your tyranny has overshadowed <strong>and</strong> transcended that <strong>of</strong><br />

government- yet you are not elected.<br />

You have highjacked our space, our freedom <strong>and</strong> our sovereignty.<br />

You have no moral right to rule us or possess us.<br />

You are not welcome among us.<br />

You do not answer to democracy, but you will have to answer to<br />

us, <strong>the</strong> People.<br />

Your arrogance <strong>and</strong> indifference to human tragedy, your abuse<br />

<strong>and</strong> avid concentration <strong>of</strong> power <strong>and</strong> hegemony disqualify you<br />

from <strong>the</strong> role you have usurped.<br />

You have transformed our dreams <strong>and</strong> promises <strong>of</strong> decentralised<br />

freedom into a centralised nightmare.<br />

Dan Shefet: Lawyer at <strong>the</strong> Paris Court <strong>of</strong> Appeal<br />

(France)<br />

Dan Shefet is a frequent speaker at international<br />

conferences <strong>and</strong> academic venues on IT Law, Data<br />

Privacy Content Regulation <strong>and</strong> Human Rights on<br />

<strong>the</strong> Internet. In 2014, he founded <strong>the</strong> Association<br />

for Accountability <strong>and</strong> Internet Democracy (AAID),<br />

<strong>the</strong> main objective <strong>of</strong> which is to introduce a general<br />

principle <strong>of</strong> accountability on <strong>the</strong> internet.<br />

Under <strong>the</strong> pretext <strong>of</strong> saving <strong>the</strong> world you fill your c<strong>of</strong>fers<br />

without contributing to society.<br />

You yield before no manipulation.<br />

You have degraded us to products in your realm. Products that<br />

deserve no respect.<br />

Governments cannot <strong>and</strong> will not stop you, but we <strong>the</strong> People will.<br />

We will boycott your services. We will take our clicks away from<br />

you <strong>and</strong> your empire will crumble.<br />

You have no right to make judgments affecting life <strong>and</strong> death.<br />

We no longer trust you.<br />

What you fear most is our passion. What you lack most is<br />

compassion.<br />

Privacy is not dead. Get over it.<br />

Now, here are our dem<strong>and</strong>s:<br />

You will give us back control over our lives.<br />

We will only allow you to co-exist with us if you accept that with<br />

power comes obligations <strong>and</strong> that your disregard <strong>of</strong> ethics <strong>and</strong><br />

social responsibility will cease.<br />

You are not above <strong>the</strong> law <strong>and</strong> your failure to assume your<br />

obligations <strong>and</strong> contribute towards society are shameful.<br />

Your abuse <strong>of</strong> your overwhelmingly dominant position <strong>and</strong> your<br />

failure to take ethics into account when developing, or ra<strong>the</strong>r,<br />

acquiring new technology <strong>and</strong> services (you develop little yourself)<br />

are intolerable.<br />

You will allow Data Subjects all over <strong>the</strong> world to take<br />

control <strong>of</strong> <strong>the</strong>ir own data <strong>and</strong> immediately heed <strong>the</strong>ir<br />

dem<strong>and</strong>s for deletion <strong>and</strong> de-referencing.<br />

You will not anoint yourself to be <strong>the</strong> judge over <strong>the</strong><br />

appropriateness <strong>of</strong> such dem<strong>and</strong>s.<br />

You will allow public inspection to verify that <strong>the</strong>se dem<strong>and</strong>s<br />

have been met <strong>and</strong> that data has been definitely <strong>and</strong> permanently<br />

erased from your gigantic data troves - because we do not trust<br />

you.<br />

You will create an ethics board with independent experts who will<br />

monitor all <strong>of</strong> your new “technology” whe<strong>the</strong>r developed in-house<br />

or acquired <strong>and</strong> in particular those technologies that apply to<br />

artificial intelligence.<br />

You will refrain from fur<strong>the</strong>r abuse <strong>of</strong> your dominant position<br />

destroying any alternative to your « services » <strong>and</strong> consolidating<br />

your power even more.<br />

This time we will create a true civilization <strong>of</strong> <strong>the</strong> Mind in<br />

Cyberspace.<br />

May it be more humane <strong>and</strong> fair than <strong>the</strong> world you have made.<br />

P a r i s, France<br />

August 3, 2018<br />

Dan Shefet<br />

ISSUE • 9<br />

2 3

opinion<br />

Dreams <strong>and</strong> Reality:<br />


How AI will Change<br />

<strong>Education</strong><br />

Harri Ketamo, founder <strong>of</strong> Headai, senior fellow in University <strong>of</strong> Turku<br />

Because AI will change everything, it<br />

is said that AI will change <strong>the</strong> way we<br />

learn. This couldn’t be a more misleading<br />

statement.<br />

What is learning as a phenomenon<br />

Harri Ketamo, Ph.D. is an entrepreneur with 20<br />

years experience in learning sciences <strong>and</strong> artificial<br />

intelligence as science <strong>and</strong> as business. Currently<br />

he is founder & chairman <strong>of</strong> Headai, a jobTech<br />

company solving global talent mismatch <strong>and</strong><br />

workforce re-skilling challenges with AI. He’s<br />

also actively participating in academic research<br />

as an adjunct pr<strong>of</strong>essor at Tampere University <strong>of</strong><br />

Technology <strong>and</strong> as a senior fellow at University <strong>of</strong><br />

Turku. Previously he has been founder & CEO <strong>of</strong><br />

gameMiner (game AI that learn from players), xTask<br />

(adaptive learning platform for mobile devices) <strong>and</strong><br />

SkillPixels (elementary school as a mobile game).<br />

Ketamo was awarded <strong>the</strong> Eisenhower Fellowship in<br />

2017.<br />

<strong>The</strong> way people learn is an outcome <strong>of</strong> millions <strong>of</strong> years <strong>of</strong><br />

evolution <strong>and</strong> it cannot be changed with technology. In short,<br />

we learn through interaction by observing <strong>the</strong> world around<br />

us i.e. we make observations, conceptualise our observations<br />

<strong>and</strong> connect <strong>the</strong>m into our existing underst<strong>and</strong>ing. When we<br />

connect observations we have made before, we just streng<strong>the</strong>n<br />

our current underst<strong>and</strong>ing. When adding new observations, we<br />

learn. When we make observations that don’t fit into our existing<br />

underst<strong>and</strong>ing, we might end up learning <strong>and</strong> at <strong>the</strong> same time<br />

changing our underst<strong>and</strong>ing radically.<br />

In science, <strong>the</strong> learning process is studied at <strong>the</strong> for example<br />

chemistry level, biological level, neural cell level, social level<br />

<strong>and</strong> psychological level. All science supports this generalised<br />

definition <strong>of</strong> learning, no matter if we call it conceptual learning,<br />

constructivism or connectivism. Claims that AI will change <strong>the</strong><br />

way we learn is just like a claim that television will change <strong>the</strong><br />

way we learn. Every ten years, <strong>the</strong>re have been technologies that<br />

have claimed to change <strong>the</strong> way learning happens. Games (2010),<br />

mobile devices (2000), Internet (1990), CD-rom (1980), television<br />

(1970), programmed instruction (1960), radio (1950) <strong>and</strong> spirit<br />

duplicators (1930), just to mention a few.<br />

We dream that technology would make learning fast <strong>and</strong> easy, but<br />

that’s an illusion.<br />

However, all technologies, including artificial intelligence,<br />

have changed our abilities to make observations <strong>of</strong> <strong>the</strong> world.<br />

Fur<strong>the</strong>rmore, technologies have enabled a remarkable change<br />

in education <strong>and</strong> teaching. Radio brought access to good quality<br />

educational content for millions. Internet provided access to a<br />

lot <strong>of</strong> information, readily available for learners. A decade later,<br />

<strong>the</strong>re was no need to just deliver information in schools – new<br />

pedagogical innovations were developed. <strong>The</strong> Learning process,<br />

however, remained <strong>the</strong> same: we yet connected our observations<br />

into our mental underst<strong>and</strong>ing. Learning is hard work, but we<br />

can make it more enjoyable, accessible <strong>and</strong> effective with good<br />

pedagogy. Technology alone is never <strong>the</strong> answer.<br />

Coming back to <strong>Artificial</strong> <strong>Intelligence</strong> (AI)<br />

We should not spend too much time on defining what AI is.<br />

All machine performed activities that would require conscious<br />

thinking also from people, can be called <strong>Artificial</strong> <strong>Intelligence</strong>. AI<br />

is not dependent on which algorithms or technological frameworks<br />

are used.<br />

Fur<strong>the</strong>rmore, AI is old. As a science, it comes from <strong>the</strong> 50’s <strong>and</strong> it<br />

has been with us for more than 60 years. <strong>The</strong> first generation AIs<br />

were decision trees <strong>and</strong> state machines, based on programmed<br />

rules. In many solutions, <strong>the</strong> rules were based on scientific work<br />

<strong>and</strong> millions <strong>of</strong> lines <strong>of</strong> data, but <strong>the</strong> rules were programmed<br />

after people found <strong>the</strong> rules. In second generation AI, <strong>the</strong>re are<br />

no pre-programmed rules <strong>and</strong> machines would learn <strong>the</strong> rules<br />

from <strong>the</strong> given data. In o<strong>the</strong>r words, <strong>the</strong> rules with which AI works<br />

are not programmed, <strong>the</strong>y are (machine) learned. <strong>The</strong>re are<br />

excellent use cases for both types <strong>of</strong> AI, for some cases<br />

programmed rules are perfect, because <strong>the</strong>y are fast to<br />

perform, e.g. spell checking. Some o<strong>the</strong>r cases require<br />

machine learning because <strong>of</strong> <strong>the</strong> complexity <strong>of</strong> <strong>the</strong> case,<br />

e.g. speech recognition. <strong>The</strong> third generation is going to<br />

ISSUE • 9<br />

2 5

opinion<br />

be about extending <strong>the</strong> concept <strong>of</strong> machine learning. i.e.<br />

teaching machines case by case with natural language.<br />

AI can be used as an assistant for people, as a closely managed<br />

worker or as an autonomous system. In many cases, we use AI<br />

in <strong>the</strong> form <strong>of</strong> a co-worker, letting it to do easy, boring, dirty or<br />

dangerous tasks. Assisting AI performs numerous supporting tasks<br />

for pr<strong>of</strong>essionals <strong>and</strong> so enabling <strong>the</strong>m to focus on higher order<br />

thinking. Currently, <strong>the</strong>re are very limited number <strong>of</strong> autonomous<br />

Alongside AI ethics, we should discuss<br />

data ethics: who is allowed to collect data,<br />

who owns <strong>the</strong> data, is <strong>the</strong> data real <strong>and</strong><br />

valid, what are biases in <strong>the</strong> data, what is<br />

<strong>the</strong> role <strong>of</strong> manipulation, <strong>and</strong> so on.<br />

AIs in use, however, when we refer to AI, we tend to think<br />

autonomous AI. It is important to note that <strong>the</strong> ethical discussion<br />

around AI is very different if we talk about autonomous systems<br />

or systems that provide assistance by performing routine tasks.<br />

Alongside AI ethics, we should discuss data ethics: who<br />

is allowed to collect data, who owns <strong>the</strong> data, is <strong>the</strong> data<br />

real <strong>and</strong> valid, what are biases in <strong>the</strong> data, what is <strong>the</strong><br />

role <strong>of</strong> manipulation, <strong>and</strong> so on. We have to accept we can<br />

not talk AI as an isolated system, it is always connected to o<strong>the</strong>r<br />

technologies <strong>and</strong> processes.<br />

And how AI will change education<br />

<strong>The</strong>re is nothing new in applying AI to learning. In 1912, Edward<br />

Thorndike introduced his idea about <strong>The</strong> Learning Machine. <strong>The</strong><br />

machine could ask questions <strong>of</strong> <strong>the</strong> learner <strong>and</strong> suggest fur<strong>the</strong>r<br />

readings, if required. Because <strong>the</strong>re were no modern computers<br />

invented at that point, <strong>The</strong> Learning Machine applied punch<br />

cards to run Adaptive Learning features 100 years before <strong>the</strong><br />

concept <strong>of</strong> Adaptive Learning became popular in practice.<br />

Adaptive learning applies computational methods that fit into <strong>the</strong><br />

AI-family, <strong>and</strong> so we can call it AI. <strong>The</strong>re are also two generations<br />

<strong>of</strong> adaptive learning. Prior to <strong>the</strong> 1990s, adaptive learning was<br />

based on rules found in scientific research <strong>and</strong> <strong>the</strong>n programmed<br />

into machines. Post <strong>the</strong> 90’s, real-time data was used in order to<br />

provide highly individual learning experiences for <strong>the</strong> learner <strong>and</strong><br />

detailed learning analytics. This required o<strong>the</strong>r technologies<br />

like fast internet, big memories, own devices <strong>and</strong><br />

computation power to support AI.<br />

<strong>The</strong>re are parallel opportunities to apply AI in several educational<br />

tasks to ei<strong>the</strong>r improve or speed up <strong>the</strong> process. However, <strong>the</strong><br />

power <strong>of</strong> AI is not in improving old process, we have to rethink <strong>the</strong><br />

processes.<br />

If we want to underst<strong>and</strong> <strong>the</strong> major opportunities related to AI,<br />

we have to focus on one <strong>of</strong> <strong>the</strong> biggest global challenges we face in<br />

education: We don’t have enough teachers to provide education for<br />

everyone. This is critical in elementary education, but remarkable<br />

also in vocational training, higher education <strong>and</strong> lifelong learning.<br />

In early years <strong>of</strong> education, such as elementary education <strong>and</strong><br />

primary education, <strong>the</strong> human teacher cannot be replaced. No<br />

way. S<strong>of</strong>t skills <strong>and</strong> transferrable skills such as critical thinking,<br />

problem solving, team work <strong>and</strong> communication needs interaction<br />

with people. However, we can provide AI-based teaching assistants<br />

<strong>and</strong> helpers for teachers <strong>and</strong> so enable teachers to work with<br />

bigger groups. This, however, requires AI-pedagogy, which we<br />

don’t have it yet. We need to invest a lot in pedagogical research<br />

in order to underst<strong>and</strong> <strong>the</strong> full potential <strong>of</strong> AI to elementary<br />

education.<br />

In lifelong learning, work place training <strong>and</strong> pr<strong>of</strong>essional<br />

development, <strong>the</strong> key factor is to underst<strong>and</strong> what skills are needed<br />

in <strong>the</strong> labour markets now <strong>and</strong> in <strong>the</strong> future. Although <strong>the</strong>re is<br />

research being undertaken, big data is not being applied <strong>and</strong><br />

micro-skills are not being focused upon because <strong>the</strong>re is too much<br />

data <strong>and</strong> complexity for people.<br />

For a couple <strong>of</strong> years, we have had AI assistance that reads<br />

through terabytes <strong>of</strong> job openings, curriculums, future forecast<br />

reports <strong>and</strong> labor market analysis, to build a real time research<br />

report on skills seeking. We could do this also with people, if we<br />

just could have millions <strong>of</strong> people to do this work.<br />

AI can build parallel models on an individual’s skills at micro<br />

level, it can easily show an individualised career path: direct jobs<br />

with person’s current skills as well as low-hanging jobs that are<br />

accessible after person gets couple <strong>of</strong> identified skills. Fur<strong>the</strong>rmore,<br />

when AI knows <strong>the</strong> curriculums, it can also reveal educational<br />

opportunities relevant for individual.<br />

When <strong>the</strong> skill gap is recognised, AI can construct tests <strong>and</strong><br />

curate content packages for <strong>the</strong> individual. In o<strong>the</strong>r words, AI<br />

can produce online courses from identified skills <strong>and</strong> so bring<br />

vocational training <strong>and</strong> lifelong learning for billions <strong>of</strong> learners.<br />

Finally, AI won’t change <strong>the</strong> way we learn, but it can change <strong>the</strong><br />

access to education, globally.<br />

Download #YouthWagingPeace now.<br />

Also available, Action Guidelines for Prevention <strong>of</strong> Violent Extremism (English <strong>and</strong> French versions)<br />

Details, visit : http://mgiep.unesco.org/youth-waging-peace<br />


International Youth Campaign<br />

on Kindness for <strong>the</strong> SDGs<br />

1<br />

About<br />

<strong>The</strong> UNESCO MGIEP has initiated <strong>the</strong> International Youth Campaign on Kindness for <strong>the</strong> Sustainable Development Goals (SDGs).<br />

<strong>The</strong> purpose <strong>of</strong> this Campaign is to mobilise <strong>the</strong> world’s youth to achieve <strong>the</strong> 17 SDGs through transformative acts <strong>of</strong> kindness. <strong>The</strong><br />

overall goals <strong>of</strong> <strong>the</strong> campaign are to:<br />

2<br />

• Create a positive culture <strong>of</strong> kindness where a person’s selfless act matters <strong>and</strong> to show how it can make a difference to <strong>the</strong> SDGs<br />

• Provide a comprehensive, independent, fun <strong>and</strong> engaging pathway for youth to demonstrate how an act <strong>of</strong> kindness is a non-trivial<br />

task <strong>and</strong> that being kind is “cool”<br />

3<br />

4<br />

5<br />




http://mgiep.unesco.org/kindness<br />

1. Bhubaneshwar City, India: SDG 9: Innovation, Industry <strong>and</strong> Infrastructure<br />

2 & 3. Kolkata City, India: SDG 4: Quality <strong>Education</strong>: An educational session at two NGO<br />

4. Lucknow City, India: SDG 2: Zero Hunger<br />

5. Kolkata City, India: SDG 4: Quality <strong>Education</strong>: An educational session at two NGO’s<br />

6. Lucknow City, India: SDG 2: Zero Hunger: Food distribution among <strong>the</strong> underprivileged in<br />

Lucknow (in collaboration with <strong>the</strong> Robin Hood Army)<br />

6<br />

2 9

cover<br />

story<br />


Personalising ‘Learning’ -<br />

Can AI Promise<br />

Customised <strong>Education</strong><br />

For ‘Humanity’<br />

N<strong>and</strong>ini Chatterjee Singh<br />

Programme Specialist - Science Of Learning, Unesco Mgiep<br />

Raunak Jain<br />

Former Consultant, UNESCO MGIEP<br />

B<br />

outique or personalised<br />

learning is <strong>the</strong> final<br />

frontier <strong>of</strong> a successful<br />

education system.<br />

Recent research in neuroscience has<br />

demonstrated that each brain is ‘wired<br />

uniquely’ <strong>and</strong> thus ‘learns differently’.<br />

Thus far, education systems have focused<br />

on curricula geared towards producing a<br />

workforce for mass industrial deployment<br />

<strong>and</strong> have implemented <strong>the</strong> ‘one size fits<br />

all’ approach. Unfortunately, this has led<br />

to <strong>the</strong> following undesirable results:<br />

(1) Many individuals trained but unable to<br />

adapt to changing job requirements,<br />

(2) Little consideration for natural talent,<br />

<strong>the</strong>reby killing creativity,<br />

(3) Little or no love for learning,<br />

(4) Unhappy, frustrated youth.<br />

Consequently, <strong>the</strong> current education<br />

systems finds itself broken, insufficient<br />

<strong>and</strong> ineffective at meeting <strong>the</strong> dem<strong>and</strong>s <strong>of</strong><br />

<strong>the</strong> 21 st century.<br />

For education to provide holistic<br />

development to children, it is urgent <strong>and</strong><br />

necessary that learning become a fun<br />

<strong>The</strong> resurgence <strong>of</strong><br />

<strong>Artificial</strong> <strong>Intelligence</strong> or<br />

AI, however, <strong>of</strong>fers a ray<br />

<strong>of</strong> hope in achieving such<br />

a ‘personalised learning<br />

system<br />

<strong>and</strong> enjoyable experience. It should allow<br />

children to learn at <strong>the</strong>ir own pace, it<br />

should do so in ways that address <strong>the</strong>ir<br />

various strengths <strong>and</strong> weaknesses, <strong>and</strong><br />

most importantly, it should prepare <strong>the</strong>m<br />

to realise <strong>the</strong>ir full potential. For this<br />

to happen, learning needs to become<br />

‘personal’. While <strong>the</strong> premise is exciting<br />

<strong>and</strong> promising, such a personalised<br />

learning system has seemed utopian thus<br />

far.<br />

<strong>The</strong> resurgence <strong>of</strong> <strong>Artificial</strong><br />

<strong>Intelligence</strong> or AI, however, <strong>of</strong>fers<br />

a ray <strong>of</strong> hope in achieving such a<br />

‘personalised learning system’. In<br />

order to underst<strong>and</strong> how AI can make<br />

this possible, we will give an overview <strong>of</strong><br />

AI, how it has resurfaced, examples <strong>of</strong> its<br />

application, followed by how we expect AI<br />

to facilitate ‘personalised learning’.<br />

What is AI ?<br />

<strong>The</strong> term <strong>Artificial</strong> <strong>Intelligence</strong> or<br />

AI was coined by John McCarthy in<br />

1956, two years after <strong>the</strong> untimely<br />

death <strong>of</strong> Alan Turing, who came<br />

to be known as <strong>the</strong> fa<strong>the</strong>r <strong>of</strong> AI. In<br />

1950, at a time when <strong>the</strong> first generalpurpose<br />

computers were being built,<br />

Turing was already grappling with <strong>the</strong><br />

question “Can machines think?” He<br />

developed a hypo<strong>the</strong>tical machine,<br />

called a Turing machine, for encrypting<br />

codes built to test <strong>and</strong> define Machine<br />

<strong>Intelligence</strong> <strong>and</strong> thus refers to computer<br />

programs that exhibit human-like<br />

ISSUE • 9<br />

3 1


intelligence such as logical reasoning,<br />

problem solving <strong>and</strong> 1 learning.<br />

<strong>The</strong> Turing machine was able to<br />

implement algorithms that could mimic<br />

<strong>the</strong> process <strong>of</strong> basic human decisionmaking.<br />

A number <strong>of</strong> algorithms were<br />

developed <strong>and</strong> some <strong>of</strong> <strong>the</strong>m were<br />

implemented successfully in cracking<br />

secret codes during <strong>the</strong> war; however,<br />

many algorithms remain untested due to<br />

<strong>the</strong> limitations <strong>of</strong> computing capabilities.<br />

With recent advances in both computing<br />

power <strong>and</strong> algorithmic efficiency, AI<br />

gained traction. This trend is expected to<br />

accelerate as new computing technologies<br />

such as neural processors <strong>and</strong> quantum<br />

computing achieve stability <strong>and</strong> maturity.<br />

AI technology is projected to dramatically<br />

change <strong>the</strong> types <strong>of</strong> tasks performed<br />

by humans <strong>and</strong> those performed by<br />

machines. As a consequence, AI has<br />

caught <strong>the</strong> imagination <strong>of</strong> not only<br />

technologists <strong>and</strong> entrepreneurs, but<br />

also educationists <strong>and</strong> policy makers. AI<br />

is unique in being a generic umbrella<br />

technology that can enable many different<br />

disruptive technologies <strong>and</strong> solutions.<br />

Creating a Framework for<br />

AI<br />

<strong>The</strong> conception <strong>of</strong> an AI system indicates<br />

that it should possess <strong>the</strong> ability to interact<br />

with <strong>the</strong> real world. To underst<strong>and</strong> AI’s<br />

progress thus far, we will elaborate on <strong>the</strong><br />

four dimensions identified by DARPA<br />

( 2 Defence Advanced Research Projects<br />

Agency) to underst<strong>and</strong> <strong>the</strong> evolution <strong>of</strong><br />

AI. Based on this, intelligence may be<br />

broadly be 3 described.<br />

• Perception – <strong>The</strong> ability to use<br />

sensory systems <strong>and</strong> language to obtain<br />

information about <strong>the</strong> natural world<br />

• Problem analysis – <strong>The</strong> ability to process<br />

<strong>the</strong> information obtained above in order<br />

to identify a problem, analyse it <strong>and</strong><br />

provide a solution<br />

• Abstract knowledge – <strong>The</strong> ability to<br />

abstract knowledge from perception in<br />

one domain <strong>and</strong> apply it to ano<strong>the</strong>r<br />

• Experiential learning – <strong>The</strong> ability<br />

to continually learn from real-world<br />

data patterns <strong>and</strong> refine perception <strong>and</strong><br />

knowledge<br />

Next, we will review <strong>the</strong> evolution <strong>of</strong> AI<br />

systems within <strong>the</strong> framework described<br />

above. A historical review <strong>of</strong> <strong>the</strong> evolution<br />

<strong>of</strong> AI systems reveals that <strong>the</strong>y appear to<br />

have followed a different process – one<br />

that has been driven by algorithm <strong>and</strong><br />

technology development. Based on <strong>the</strong>ir<br />

capabilities, developed AI systems have<br />

been categorised in distinct generations.<br />

Gen 1 AI – Rule-driven<br />

Learning<br />

<strong>The</strong> first generation <strong>of</strong> AI programmes<br />

described knowledge through <strong>the</strong><br />

definition <strong>of</strong> static rules. One <strong>of</strong> <strong>the</strong><br />

earliest programmes in this series was<br />

designed to played chess. Computer<br />

systems used a set <strong>of</strong> defined rules to play<br />

chess with humans <strong>and</strong> <strong>the</strong>ir algorithms<br />

became so good that <strong>the</strong> systems even<br />

beat humans. Such systems <strong>of</strong>fered limited<br />

perception <strong>of</strong> <strong>the</strong> outside world, namely<br />

that <strong>of</strong> <strong>the</strong> chessboard.<br />

Global Positioning Systems (GPS)<br />

are a second example <strong>of</strong> firstgeneration<br />

AI. <strong>The</strong>y demonstrate<br />

<strong>the</strong> ability to identify location,<br />

analyse <strong>and</strong> contextualise it on a<br />

map, <strong>and</strong> provide solutions in <strong>the</strong><br />

form <strong>of</strong> directions. GPS systems<br />

possess very limited abilities to cope<br />

with deviations from <strong>the</strong> routine, yet<br />

<strong>the</strong>ir presence <strong>and</strong> utility in our daily<br />

lives cannot be denied or ignored.<br />

When viewed through <strong>the</strong> dimensions<br />

<strong>of</strong> intelligence described above, firstgeneration<br />

AI systems are only capable<br />

<strong>of</strong> using ‘problem analysis’ for limited<br />

applications <strong>of</strong> <strong>the</strong> outside world. For<br />

instance, users <strong>of</strong> Google Maps will know<br />

that based on a set <strong>of</strong> pre-defined rules, a<br />

map prescribes a specific route. However,<br />

based on <strong>the</strong>ir experience, a driver may<br />

take a different route each day, while <strong>the</strong><br />

application is unable to learn <strong>the</strong> route<br />

taken by <strong>the</strong> driver. Thus, while firstgeneration<br />

AI computer programs possess<br />

great capabilities <strong>and</strong> are constantly being<br />

deployed in many industries for <strong>the</strong> value<br />

that <strong>the</strong>y provide, <strong>the</strong>y are capable <strong>of</strong> only<br />

implementing rules <strong>and</strong> are unable to<br />

learn from patterns <strong>and</strong> implement new<br />

trajectories or outputs.<br />

Gen 2 AI – Data-driven<br />

Learning<br />

<strong>The</strong> next big leap in AI systems happened<br />

with <strong>the</strong> advent <strong>of</strong> Big Data. For a<br />

long time, access to limited data, which<br />

was isolated <strong>and</strong> held in data cocoons,<br />

prevented AI from being tested or<br />

implemented in demographic settings.<br />

Recent advances in data collection<br />

<strong>and</strong> assimilation techniques have made<br />

available to us enormous amounts <strong>of</strong><br />

data, which is analysed using advanced<br />

data analytics methods – this entire<br />

process is referred to as Big Data. Fur<strong>the</strong>r,<br />

<strong>the</strong> migration from regular CPUs to<br />

extremely powerful computing devices<br />

has enabled <strong>the</strong> development <strong>of</strong> parallel<br />

intelligence engines. AI programmes have<br />

leveraged this to great advantage; for<br />

instance, novel transformative algorithms<br />

are now available to identify clusters<br />

<strong>and</strong> patterns in datasets. Often, patterns<br />

not visible to humans can be located<br />

using such algorithms, sometimes even<br />

through trial <strong>and</strong> error. Thus, training<br />

a system using Big Data has made a<br />

huge impact on utilising <strong>the</strong> power <strong>of</strong><br />

‘pattern recognition’. However, it is brittle<br />

– it breaks as soon as soon <strong>the</strong> data set<br />

changes.<br />

One example <strong>of</strong> second-generation AI’s<br />

implementation is Micros<strong>of</strong>t’s Twitter<br />

Bot 9 <strong>and</strong> <strong>the</strong> controversy surrounding it.<br />

Popularly named Tay, this teen-talking AI<br />

chatbot was built to mimic <strong>and</strong> converse<br />

with users in real time; it was released via<br />

Twitter in 2016. As it began to converse<br />

with users, some people took advantage<br />

<strong>of</strong> Tay’s machine learning capabilities<br />

<strong>and</strong> began to teach it racist <strong>and</strong> <strong>of</strong>fensive<br />

language. As a consequence, <strong>the</strong> bot<br />

began to post <strong>of</strong>fensive tweets through its<br />

Twitter account, forcing Micros<strong>of</strong>t to shut<br />

down <strong>the</strong> service only 16 hours after its<br />

launch.<br />

Categories <strong>of</strong><br />

<strong>Artificial</strong> <strong>Intelligence</strong><br />

Gen 1 AI –<br />

Rule Driven Learning<br />

<strong>The</strong> first generation <strong>of</strong> AI programs that were<br />

developed incorporated computer programs that<br />

described knowledge through <strong>the</strong> definition <strong>of</strong><br />

static rules. One <strong>of</strong> <strong>the</strong> earliest in this series, was<br />

<strong>the</strong> development <strong>of</strong> computer programs that<br />

played chess. Such systems <strong>of</strong>fered limited<br />

perception <strong>of</strong> <strong>the</strong> outside world, namely that <strong>of</strong><br />

<strong>the</strong> chessboard.<br />

Gen 2 AI –<br />

Data Driven Learning<br />

<strong>The</strong> next big leap in AI systems happened with<br />

<strong>the</strong> collection <strong>and</strong> availability <strong>of</strong> BigData. For a<br />

long time, access to small limited data, which<br />

was isolated <strong>and</strong> held in data cocoons, prevented<br />

AI from being tested or implemented in<br />

population settings<br />

Gen 3 AI –<br />

Context Driven Learning<br />

Building from <strong>the</strong> earlier paragraph, it this<br />

<strong>the</strong>refore quite clear that <strong>the</strong> primary focus <strong>of</strong><br />

3rd Gen AI is to build <strong>and</strong> incorporate algorithms<br />

that underst<strong>and</strong> context <strong>and</strong> implement an<br />

analytical path to decide response.<br />


ISSUE • 9<br />

1 http://cpsc.yale.edu/research/artificial-intelligence<br />

2<br />

https://www.darpa.mil/about-us/darpa-perspective-on-ai<br />

3<br />

https://www.iata.org/publications/Documents/AI-White-Paper.pdf<br />

3 3


While <strong>the</strong> incorporation <strong>of</strong> Big Data<br />

<strong>of</strong>fers tremendous promise <strong>and</strong> is exciting,<br />

second-generation AI lacks <strong>the</strong> power <strong>of</strong><br />

logical reasoning, underst<strong>and</strong>ing context<br />

<strong>and</strong> abstracting knowledge to different<br />

domains. Human beings learn in multiple<br />

ways – through experience, patterns,<br />

<strong>and</strong> imitation – to list a few. Clearly,<br />

such ideas needed to be translated to<br />

algorithms <strong>and</strong> incorporated in thirdgeneration<br />

AI.<br />

it would need to underst<strong>and</strong> context<br />

<strong>and</strong> have <strong>the</strong> capability to go from one<br />

domain to ano<strong>the</strong>r; abstracting knowledge<br />

<strong>and</strong> learning. Thus, <strong>the</strong> possibility <strong>of</strong> a<br />

system or machine underst<strong>and</strong>ing <strong>the</strong><br />

context <strong>of</strong> its environment, <strong>and</strong> learning<br />

<strong>and</strong> adapting based on changes in that<br />

environment, is possibly <strong>the</strong> next frontier<br />

<strong>of</strong> AI.<br />

<strong>The</strong> Role <strong>of</strong> AI in <strong>Education</strong><br />

To underst<strong>and</strong> <strong>the</strong> role <strong>of</strong> AI<br />

in education, we first have<br />

to underst<strong>and</strong> <strong>the</strong> concepts<br />

which make education<br />

systems successful, which<br />

will <strong>the</strong>n help us underst<strong>and</strong><br />

how AI can add value to<br />

each component in <strong>the</strong><br />

system.<br />


Gen 3 AI – Context-driven<br />

Analysis<br />

Building upon our argument, it is quite<br />

clear that <strong>the</strong> primary focus <strong>of</strong> Gen 3 AI<br />

is to build <strong>and</strong> incorporate algorithms<br />

that underst<strong>and</strong> context <strong>and</strong> implement<br />

an analytical path to decide response. As<br />

discussed earlier, <strong>the</strong> data sets on which AI<br />

systems are trained are brittle. Thus, <strong>the</strong><br />

Gen 2 AI is only as good as its data set it is<br />

trained on <strong>and</strong> is incapable <strong>of</strong> recognising<br />

right from wrong. For instance, as seen<br />

in <strong>the</strong> case <strong>of</strong> Tay, when <strong>the</strong> bot was<br />

provided new patterns by Twitter users,<br />

it learnt new patterns, not realising that<br />

<strong>the</strong>se were inappropriate patterns or even<br />

different from what it was taught earlier.<br />

This brings to <strong>the</strong> fore <strong>the</strong> value <strong>of</strong> logical<br />

reasoning <strong>and</strong> analysis in AI systems.<br />

<strong>The</strong> second aspect that Gen 3 AI needs<br />

to incorporate is recognising patterns<br />

across domains <strong>and</strong> abstractions. Humans<br />

beings’ ability to not only recognise<br />

patterns but abstract knowledge<br />

such that it can be used in different<br />

settings is incredibly powerful <strong>and</strong><br />

is learnt <strong>and</strong> accomplished without too<br />

much data <strong>and</strong> quite early in life. One<br />

simple example is our ability to recognise<br />

h<strong>and</strong>writing, which varies hugely among<br />

people. Yet, <strong>the</strong> human brain is capable<br />

<strong>of</strong> recognising a word written in various<br />

ways. One faculty that <strong>the</strong> human brain<br />

uses here is <strong>the</strong> underst<strong>and</strong>ing <strong>of</strong> context,<br />

which is a vital aspect <strong>the</strong> way we make<br />

decisions. For AI systems to be more<br />

effective, it is important that <strong>the</strong>y be<br />

able to take decisions based on context.<br />

If third-generation AI is to be effective,<br />

To underst<strong>and</strong> <strong>the</strong> role <strong>of</strong> AI<br />

in education, we must first<br />

underst<strong>and</strong> <strong>the</strong> concepts that make<br />

education systems successful,<br />

which will <strong>the</strong>n help us underst<strong>and</strong><br />

how AI can add value to each<br />

component in <strong>the</strong> system. <strong>The</strong> two<br />

primary components <strong>of</strong> any education<br />

system are <strong>the</strong> student <strong>and</strong> <strong>the</strong> teacher.<br />

An ideal education system should provide<br />

content that is contextually relevant,<br />

technically sound, <strong>and</strong> emotionally<br />

rewarding to ensure that students acquires<br />

skills <strong>and</strong> continue to remain motivated<br />

to 4 learn. This can be achieved if <strong>the</strong>y<br />

are provided an educational experience<br />

that is immersive, multi-sensory, paced<br />

appropriately, <strong>and</strong> rewarding.<br />

For teachers to effectively<br />

implement good teaching practices,<br />

education also needs to be engaging<br />

<strong>and</strong> fun, <strong>and</strong> most importantly,<br />

allow <strong>the</strong> teacher to continuously<br />

measure learning <strong>and</strong> receive<br />

transparent feedback.<br />

<strong>Artificial</strong> <strong>Intelligence</strong> embedded in<br />

digital learning ecosystems presents<br />

<strong>the</strong> possibility <strong>of</strong> achieving both <strong>of</strong> <strong>the</strong><br />

above in a measurable <strong>and</strong> scientific<br />

manner. It is <strong>the</strong> <strong>the</strong>oretical basis on<br />

which personalised learning <strong>and</strong> analytics<br />

work to enable richer underst<strong>and</strong>ing<br />

<strong>of</strong> learning data <strong>and</strong> improve learning<br />

effectiveness at scale.<br />

<strong>The</strong> integration <strong>of</strong> AI into learning<br />

systems is possible because <strong>of</strong> digitisation.<br />

Here, digitisation refers to <strong>the</strong> constant<br />

interaction <strong>of</strong> <strong>the</strong> individual with <strong>the</strong><br />

machine during learning, which makes it<br />

possible for us to collect data on decision<br />

making. This data would be supremely<br />

valuable to underst<strong>and</strong> students’ learning<br />

patterns, types <strong>and</strong> pace, which in turn<br />

could be used to build <strong>and</strong> train AI<br />

systems.<br />

Imagine a digital platform presented to<br />

children as an interface where <strong>the</strong>y can<br />

speak, touch <strong>and</strong> visualise information,<br />

<strong>and</strong> engage with it. This information<br />

can be provided in visual or auditory<br />

modes, or even as a short movie. Using<br />

<strong>the</strong> powerful modality <strong>of</strong> a story, children<br />

can be engaged in a learning experience<br />

that requires interaction with a digital<br />

platform. <strong>The</strong> story itself is delivered in<br />

a format that immerses children in an<br />

experience where <strong>the</strong>y respond, interact,<br />

think critically <strong>and</strong> take decisions towards<br />

specific outcomes, while <strong>the</strong>ir interactions<br />

are recorded as a digital transaction.<br />

Navigating through this story would<br />

require children to assume a different<br />

identity – a common aspect <strong>of</strong> gaming<br />

– <strong>and</strong> <strong>the</strong>n make a choice. Throughout<br />

this process, children are provided<br />

information <strong>and</strong> learn, making both<br />

intellectual <strong>and</strong> emotional decisions, <strong>and</strong><br />

constantly questioning <strong>and</strong> reflecting on<br />

<strong>the</strong> situations presented in <strong>the</strong> story. Such<br />

a process would not only allow <strong>the</strong>m to<br />

learn in a fun <strong>and</strong> engaging manner but<br />

also permit continuous assessment <strong>of</strong> <strong>the</strong><br />

different skills that need to be acquired.<br />

<strong>The</strong> digital transactions that learners<br />

would undertake on such a platform<br />

–from conversations to assessment<br />

exercises to games – would have a digital<br />

footprint that can be used to pr<strong>of</strong>ile every user. Central to this<br />

digital curriculum would be <strong>the</strong> inclusion <strong>of</strong> skills that emphasise<br />

compassion, collaboration <strong>and</strong> co-operation among fellow<br />

participants, so that children build emotional intelligence along<br />

with intellectual intelligence.<br />

For <strong>the</strong> teacher, this technology would facilitate intelligent tutoring.<br />

<strong>The</strong> teacher would be able to analyse each child’s learning path<br />

<strong>and</strong> actively generate/select content that intervenes to improve<br />

weaknesses, <strong>and</strong> design well-informed strategies that can be tested<br />

<strong>and</strong> improved in a self-learning environment.<br />

Over a period <strong>of</strong> time, <strong>the</strong> learning patterns <strong>and</strong> teacher<br />

interventions would begin to contribute to a collective intelligence,<br />

both from student <strong>and</strong> teacher, <strong>and</strong> <strong>the</strong>n be automated by AI to<br />

multiple languages that would be available across classrooms.<br />

Here, children would also benefit from <strong>the</strong> best pedagogical<br />

practices that would have been gleaned from multiple teachers. AI<br />

would also have learnt patterns from <strong>the</strong> collective intelligence <strong>and</strong><br />

thus <strong>the</strong> data set being brittle would cease to be a concern.<br />

Such a platform has been built by <strong>the</strong> UNESCO MGIEP, titled<br />

CHI (Collective Human <strong>Intelligence</strong> - chi.buzz). This learning<br />

platform will provide a safe learning environment for learners <strong>and</strong><br />

its first version was launched at TECH 2018.<br />

We have not yet determined all <strong>the</strong> skills that today’s pre-schoolers<br />

will need in <strong>the</strong> workplace <strong>of</strong> 2030. We are hopeful, however,<br />

that such interactive platforms will cater to individual needs <strong>and</strong><br />

provide ‘personalised learning’ to develop skill sets that will allow<br />

every child an equal opportunity to realise <strong>the</strong>ir full potential.<br />

ISSUE • 9<br />

4<br />

https://www.cisco.com/c/dam/en_us/solutions/industries/docs/education/Multimodal-Learning-Through-Media.pdf<br />

3 5




CHI – A Knowledge<br />


Sharing Digital Platform<br />

Powered by Machine<br />

Learning<br />

Roy Saurabh, Chief Technology Officer, UNESCO MGIEP<br />

A<br />

rtificial <strong>Intelligence</strong> (AI) <strong>and</strong> <strong>the</strong><br />

Neurosciences have been interlinked since <strong>the</strong><br />

time <strong>the</strong> field <strong>of</strong> AI was at a nascent stage.<br />

Parallels were drawn between artificial<br />

intelligence <strong>and</strong> human intelligence, as early as 1956, in<br />

a workshop held on <strong>the</strong> campus <strong>of</strong> Dartmouth College [1]. Many<br />

<strong>of</strong> <strong>the</strong> early pioneers straddled both fields, with collaborations<br />

between <strong>the</strong> disciplines <strong>of</strong> neuroscience <strong>and</strong> artificial intelligence<br />

proving to be highly productive (Churchl<strong>and</strong> <strong>and</strong> Sejnowski, 1988,<br />

Hebb, 1949, Hinton et al., 1986, Hopfield, 1982, McCulloch<br />

<strong>and</strong> Pitts, 1943, Turing, 1950) [2]. Whe<strong>the</strong>r it is Neural Network,<br />

Reinforcement Learning or <strong>the</strong> Machine Learning training<br />

Roy Saurabh has over 13 years global experience in Technology & Strategy<br />

Consulting. Roy has a proven track record in large-scale project management,<br />

Entrepreneurship, <strong>and</strong> Technology Consulting. Previously, he was <strong>the</strong> Head <strong>of</strong><br />

Technology for <strong>the</strong> National Skill Development Mission, impacting more than 3<br />

million lives, Director for Oracle’s leading Europe Partner, Cedar Consulting UK,<br />

<strong>and</strong> Founder/CEO <strong>of</strong> Zosher. Roy has in-depth knowledge <strong>of</strong> different client/server<br />

architectures based on various frameworks, data visualisation tools, classification<br />

& regression models <strong>and</strong> o<strong>the</strong>r popular as well as advanced <strong>Artificial</strong> <strong>Intelligence</strong><br />

/ NLP <strong>and</strong> Machine Learning algorithms.<br />

models, data scientists always have had a superior human<br />

intelligence counterpart to refer to while envisioning or drawing<br />

inspiration for AI models.<br />

What makes this co-existence exciting is that both fields<br />

have some highly specialised ‘knowns’ <strong>and</strong> ‘unknowns’.<br />

When we talk about blind spots in <strong>the</strong> neurosciences, a basic<br />

in-depth underst<strong>and</strong>ing <strong>of</strong> creativity, imagination, dreams<br />

<strong>and</strong> consciousness is still lacking. As a recent DARPA project<br />

study[3] indicates, while AI has taken strides in some super<br />

specialised pockets, basic human common sense is absent in<br />

at least <strong>the</strong> current applications <strong>of</strong> AI across various fields.<br />

With this acknowledgment, we tried to leverage <strong>the</strong> strengths<br />

<strong>of</strong> both <strong>the</strong> research fields <strong>of</strong> AI <strong>and</strong> <strong>the</strong> Neurosciences in our<br />

conceptualisation process to work towards helping leaners Learn.<br />

UNESCO MGIEP’s work in <strong>the</strong> field <strong>of</strong> Differential Learning<br />

<strong>and</strong> rich data sets that have been curated for advanced model<br />

training with years <strong>of</strong> education related legacy data, has helped<br />

us organically draw learnings from one field to <strong>the</strong> o<strong>the</strong>r. One<br />

<strong>of</strong> <strong>the</strong> major learnings <strong>of</strong> such a brainstorming session led to<br />

<strong>the</strong> inception <strong>of</strong> a unique knowledge sharing platform, titled<br />

Collective Human <strong>Intelligence</strong> (We prefer <strong>the</strong><br />

Greek pronunciation /kaɪ to <strong>the</strong> Kung Fu P<strong>and</strong>a3 version).<br />

CHI, quite simply, is an AI-powered digital platform<br />

that helps extract learnings from tasks that AI does well,<br />

to help teach <strong>and</strong> consequently learn. CHI helps curriculum<br />

designers, policy makers, content developers, teachers <strong>and</strong> students<br />

rethink how knowledge <strong>and</strong> intelligence can be imparted in<br />

<strong>the</strong> 21 st century. Central to <strong>the</strong> design <strong>and</strong> development <strong>of</strong> <strong>the</strong><br />

platform was <strong>the</strong> question ‘how learners learn’.<br />

Before delving into <strong>the</strong> technical vision <strong>and</strong> features <strong>of</strong> CHI,<br />

underst<strong>and</strong>ing <strong>the</strong> ‘WHY’ behind <strong>the</strong> development <strong>of</strong> CHI is<br />

critical. <strong>The</strong> two factors that fueled CHI’s conceptualisation were:<br />

• Limitations that <strong>the</strong> current content creation mainstream<br />

platforms present, both in <strong>the</strong> form <strong>of</strong> content design tools/<br />

content delivery modes <strong>and</strong> <strong>the</strong> limited data insights that existing<br />

platforms provide for <strong>the</strong> creator community.<br />

• Ownership <strong>of</strong> <strong>the</strong> data generated by existing platforms solely lies<br />

in <strong>the</strong> h<strong>and</strong>s <strong>of</strong> for-pr<strong>of</strong>it organisations, depriving <strong>the</strong> world <strong>of</strong> <strong>the</strong><br />

invaluable data insights that could help take corrective measures<br />

for <strong>the</strong> betterment <strong>of</strong> <strong>the</strong> world.<br />

<strong>The</strong>se limitations, when thought <strong>of</strong> in <strong>the</strong> context <strong>of</strong> education<br />

<strong>and</strong> learning outcomes, become even more critical because <strong>of</strong> <strong>the</strong><br />

dynamic nature <strong>of</strong> today’s education ecosystem. Since, we are<br />

trying to prepare learners for such dynamic environments; creating<br />

engaging learner tools. In many ways, this puts into practice<br />

Howard Gardner’s multiple intelligences pedagogical framework<br />

[4].<br />

When <strong>the</strong> reasons became compelling enough, <strong>the</strong> technology<br />

requirements <strong>and</strong> long-term vision began taking shape. CHI uses<br />

<strong>the</strong> latest innovative technology to ensure maximum<br />

engagement <strong>of</strong> creators <strong>and</strong> learners as well as provide<br />

deep insights from <strong>the</strong> crowdsourced learner & creator<br />

datasets. <strong>The</strong> scale <strong>of</strong> <strong>the</strong> impact <strong>and</strong> <strong>the</strong> number <strong>of</strong> data points<br />

per platform-registered user are a dream for any Technology team<br />

ISSUE • 9<br />

3 7


trying to do something meaningful with <strong>Artificial</strong> <strong>Intelligence</strong><br />

(Deep Learning in particular). With rich Datasets to train on,<br />

<strong>the</strong> key features that we are setting out within CHI are Machine<br />

Translation [5], Sequence Tagging, Sentiment & Emotion<br />

Analysis, Recommendation Engine <strong>and</strong> H<strong>and</strong>writing Recognition.<br />

<strong>The</strong>se features, when implemented in conjunction with learnercentric<br />

platform design, would slowly start to unleash <strong>the</strong> power<br />

<strong>of</strong> technology in <strong>the</strong> educational context. <strong>The</strong> platform, in <strong>the</strong><br />

first phase, tries to solve real issues that are left unaddressed in <strong>the</strong><br />

legacy platforms from both learners’ & creators’ perspectives.<br />

From <strong>the</strong> content creator’s point <strong>of</strong> view, CHI <strong>of</strong>fers rich<br />

content creation abilities, including embedding audio,<br />

rich text, video, journaling, games (developed using<br />

Unity/Phaser) to name a few. <strong>The</strong> platform framework<br />

design has been made extremely intuitive <strong>and</strong> powerful to help<br />

significantly reduce creators’ learning curve. Creators are fur<strong>the</strong>r<br />

armed with real-time emotional analysis insights related to course<br />

reviews, discussions <strong>and</strong> facilitated dialogues. <strong>The</strong>se insights help<br />

creators identify <strong>the</strong> temotion shift on pertinent topics <strong>and</strong> identify<br />

influencers by studying context-specific trends. Fur<strong>the</strong>r, with a<br />

powerful character recognition feature, CHI helps content creators<br />

in uploading <strong>and</strong> transcribing <strong>of</strong>fline legacy content, thus reducing<br />

<strong>the</strong> data entry effort drastically. Finally, advanced analytics helps<br />

creators be mindful <strong>of</strong> what is working <strong>and</strong> what is not in terms<br />

<strong>of</strong> <strong>the</strong> learning outcomes <strong>of</strong> <strong>the</strong> respective courses. Creators<br />

have <strong>the</strong> ability to ‘create’ customised learner-centric courses by<br />

personalising features such as <strong>the</strong> layout <strong>of</strong> <strong>the</strong> learning blocks<br />

or button design. Simultaneously, <strong>the</strong>y have access to advanced<br />

analytics such as student dropout rate <strong>and</strong> time spent on lessons,<br />

modules or <strong>the</strong> courses. Predictive analytics within CHI can<br />

potentially help creators take pro-active actions, especially in <strong>the</strong><br />

case <strong>of</strong> at-risk learners who may be on <strong>the</strong> verge <strong>of</strong> dropping out.<br />

We hope that <strong>the</strong>se power-packed features <strong>of</strong> CHI are leveraged<br />

at scale, helping embed global best practices for future content<br />

creators.<br />

<strong>The</strong> real strength <strong>of</strong> CHI, however, lies with <strong>the</strong> various features<br />

available for learners; <strong>the</strong> development <strong>of</strong> CHI was conceptualised<br />

with learners at <strong>the</strong> center. Learners experience a guided yet<br />

flexible <strong>and</strong> a highly personalised learning environment. Every<br />

interaction on <strong>the</strong> platform is utilised as a way to learn more about<br />

<strong>the</strong> learning habits <strong>of</strong> <strong>the</strong> learner to help achieve <strong>the</strong> learning<br />

outcomes using <strong>the</strong> best available digital pedagogical learning<br />

mode within CHI.<br />

Fur<strong>the</strong>r, deeply embedded analytics <strong>and</strong> user behaviour indicators<br />

<strong>of</strong> learners as well as collaborative-based recommendation<br />

engines help in providing recommendations <strong>of</strong> <strong>the</strong> most suitable<br />

learning resources available from <strong>the</strong> CHI repository for learners,<br />

in areas where <strong>the</strong>y need maximum support. Learners have an<br />

opportunity to experience <strong>the</strong> immersive courses in <strong>the</strong> form<br />

that <strong>the</strong>y are most accustomed to or adept at, without diluting<br />

<strong>the</strong> course’s learning outcomes. With contextual metadata, CHI<br />

facilitates instant feedback, review as well as AI powered help<br />

facilitating trial & error learning in a judgement-free environment.<br />

CHI leverages AI for adapting to <strong>the</strong> pace <strong>of</strong> learning <strong>and</strong> can<br />

consistently <strong>of</strong>fer more complex tasks suited to <strong>the</strong> pace <strong>of</strong> <strong>the</strong><br />

learners, making learning completely non-linear. Finally, learners<br />

are provided an opportunity to experience unique content (such<br />

as course material on SEL developed by UNESCO MGIEP) that<br />

would be exclusive to CHI.<br />

We acknowledge that <strong>the</strong> sheer mention <strong>of</strong> <strong>Artificial</strong> <strong>Intelligence</strong><br />

or AI in various fields invokes mixed reactions <strong>and</strong> feelings. Some<br />

see automation in general <strong>and</strong> AI in particular as a threat to<br />

current <strong>and</strong> future jobs. Few subscribe to even more extreme views<br />

quoting Elon Musk’s dire warning about AI[6] in every context<br />

that even vaguely mentions automation. <strong>The</strong>n <strong>the</strong>re are some, who<br />

are downright excited about AI as <strong>the</strong> long-awaited game changer<br />

that is bound to make us more human by automating everything<br />

else that doesn’t fit <strong>the</strong> definition <strong>of</strong> our innate ‘humanness’. At <strong>the</strong><br />

end <strong>of</strong> <strong>the</strong> day, it does not matter which end <strong>of</strong> <strong>the</strong> spectrum you<br />

fall under on <strong>the</strong> topic <strong>of</strong> AI, one thing that no one cannot deny<br />

is that <strong>Artificial</strong> <strong>Intelligence</strong> has already made its way in almost all<br />

sectors. <strong>The</strong>refore, in our view, one <strong>of</strong> <strong>the</strong> first things that requires<br />

our immediate attention is <strong>the</strong> role <strong>of</strong> digital technology in field<br />

<strong>of</strong> education. While for most <strong>of</strong> us, our education systems did not<br />

or could not prepare us for <strong>the</strong> Fourth Industrial Revolution or<br />

what we would prefer to call as <strong>the</strong> First <strong>Intelligence</strong> Revolution, it<br />

would be remiss for us to make <strong>the</strong> same mistake by not preparing<br />

our future generations for a change that is imminent.<br />

With so much to <strong>of</strong>fer to all stakeholders, <strong>the</strong> long-term vision<br />

requires patronage from <strong>the</strong> Application Development community.<br />

CHI reach its true potential only when supported by a global<br />

community <strong>of</strong> learners, content developers <strong>and</strong> application<br />

developers. We believe that a community committed to learning<br />

will ensure CHI stays true to that vision; embedded in that vision<br />

is a truly open-source, community-supported platform that gains<br />

from <strong>the</strong> Collective Human <strong>Intelligence</strong>. It is no doubt a gr<strong>and</strong><br />

experiment but one that needs to be explored <strong>and</strong> implemented if<br />

we truly believe in transforming education for humanity!<br />

References<br />

[1] https://pdfs.semanticscholar.org/d486/9863b5da0fa4ff-<br />

5707fa972c6e1dc92474f6.pdf<br />

[2] https://www.cell.com/neuron/fulltext/S0896-6273(17)30509-<br />

3<br />

[3] https://techcrunch.com/2018/10/11/darpa-wants-to-teach<strong>and</strong>-test-common-sense-for-ai/<br />

[4] https://www.niu.edu/facdev/_pdf/guide/learning/howard_<br />

gardner_<strong>the</strong>ory_multiple_intelligences.pdf<br />

[5] https://arxiv.org/pdf/1808.03867.pdf<br />

[6] https://www.cnbc.com/2018/03/13/elon-musk-at-sxsw-a-i-ismore-dangerous-than-nuclear-weapons.html<br />

ISSUE • 9<br />

3 9

interview<br />

Interview with<br />

can also be <strong>of</strong> immense help in areas such as teacher performance<br />

analysis <strong>and</strong> teacher placements. In higher education, data science<br />

can help in student on-boarding, curriculum design <strong>and</strong> student<br />

lifecycle management.<br />

AI can also facilitate experiential learning by simulating study tours,<br />

which are typically only undertaken by students living in urban<br />

areas. Videos, augmented reality <strong>and</strong> virtual reality can make this<br />

happen, for instance, by simulating historic monuments, l<strong>and</strong>forms,<br />

motor mechanics, <strong>and</strong> more.<br />


Dr. Avik Sarkar,<br />

Head – Data Analytics Cell<br />

at NITI Aayog, Govt. <strong>of</strong> India<br />

An Interview with Akriti Mehra<br />

3. What do you perceive as threats <strong>of</strong> AI in<br />

education / how can <strong>the</strong>se be addressed?<br />

Teachers are <strong>the</strong> most important factor in ensuring <strong>the</strong> success <strong>of</strong><br />

education. <strong>The</strong> most successful universities are <strong>the</strong> ones with <strong>the</strong><br />

best teachers. In <strong>the</strong> context <strong>of</strong> schools, <strong>the</strong> countries with <strong>the</strong> best<br />

educational outcomes have a rigorous process to select teachers <strong>and</strong><br />

reward <strong>the</strong>m for imparting quality education. Replacing teachers<br />

with AI-based adaptive learning robots could lead to low learning<br />

outcomes in students. <strong>The</strong> education sector should be wary <strong>of</strong><br />

replacing teachers; instead, it should use AI to complement<br />

teachers’ efforts in imparting quality education.<br />

6. What is <strong>the</strong> current l<strong>and</strong>scape <strong>of</strong> AI <strong>and</strong><br />

education in India in terms <strong>of</strong> <strong>the</strong> no. <strong>of</strong> companies<br />

/ institutions working on this?<br />

Several e-learning platforms are gaining popularity in India.<br />

<strong>The</strong>re is a huge opportunity for <strong>the</strong>se e-learning platforms to<br />

help fill students’ learning gaps by explaining concepts in a<br />

simplified manner through <strong>the</strong> use <strong>of</strong> digital technologies. <strong>The</strong><br />

Government <strong>of</strong> India has also launched its e-learning platform to<br />

enhance <strong>the</strong> learning process for students.<br />

1. How will AI impact <strong>the</strong> future <strong>of</strong> education?<br />

<strong>Artificial</strong> <strong>Intelligence</strong> or AI provides predictions on new scenarios<br />

based on learnings from large volumes <strong>of</strong> historical data. We have<br />

to take a holistic look at a range <strong>of</strong> digital transformation in <strong>the</strong><br />

education sector, along with aspects such as early prediction <strong>of</strong><br />

student marks. Augmented reality, virtual reality <strong>and</strong> videos can<br />

provide great help to students in visualising various letarning items.<br />

Blockchain can be used to validate degree certificates assigned to<br />

students. Hence, we should look at range <strong>of</strong> digital technologies<br />

ra<strong>the</strong>r than focusing solely on AI in <strong>the</strong> education sector.<br />

Adaptive self-learning is <strong>the</strong> most popular AI application for <strong>the</strong><br />

use <strong>of</strong> data, analytics, <strong>and</strong> AI for governance, policy-making <strong>and</strong><br />

inclusive growth in India. Again, this would require personalised<br />

PCs/laptops – how many students in India can afford this? Hence,<br />

adaptive self-learning can be possibly suggested for higher studies<br />

or while preparing for entrance exams.<br />

Dr Avik Sarkar is an Officer on Special Duty at NITI Aayog, a premier policy thinktank<br />

<strong>of</strong> <strong>the</strong> Government <strong>of</strong> India. Dr Sarkar is heading <strong>the</strong> Data Analytics Cell at NITI<br />

Aayog, developing a roadmap for <strong>the</strong> use <strong>of</strong> data, analytics, <strong>and</strong> AI for governance,<br />

policy-making <strong>and</strong> inclusive growth in India. Dr Sarkar has over 17 years <strong>of</strong> experience<br />

across different aspects <strong>of</strong> data analytics, statistical modeling, data/text mining across<br />

companies such as IBM, Accenture, Nokia, NASA, Persistent Systems, Zycus, etc.<br />

He has been nominated among <strong>the</strong> “Top 10 Data Scientists in India” in 2017 by <strong>the</strong><br />

Analytics India Magazine.<br />

2. How can AI, big data <strong>and</strong> data science be used<br />

effectively in education?<br />

We need to take a step back <strong>and</strong> examine what data is<br />

available from India’s education sector to enable data science<br />

<strong>and</strong> related activities. Once data is available, it can be used for<br />

discovering patterns or predicting insights from data. For example,<br />

in India, while information on schools is collected, data on student<br />

performance is hardly available in <strong>the</strong> digital format. This results<br />

in limited potential for analysis. Access to student performance<br />

<strong>and</strong> o<strong>the</strong>r student-related data in <strong>the</strong> digital format can help<br />

predict future scores <strong>and</strong> potential drop-outs.<br />

Never<strong>the</strong>less, <strong>the</strong> DISE (District Information System for <strong>Education</strong>)<br />

has excellent data on school infrastructure, facilities, etc. Data<br />

science can help underst<strong>and</strong> patterns that are most relevant to<br />

predict student drop-outs <strong>and</strong> low learning outcomes in schools. It<br />

4. How will <strong>the</strong> role <strong>of</strong> teachers / educators<br />

be affected with <strong>the</strong> permeation <strong>of</strong> digital<br />

technologies <strong>and</strong> AI?<br />

AI <strong>and</strong> o<strong>the</strong>r digital technologies should be used to help<br />

teachers in imparting education. <strong>The</strong>y can be used for predicting<br />

student potential <strong>and</strong> marks as well as chances <strong>of</strong> at-risk students<br />

dropping out. This information would be valuable for teachers in<br />

targeting specific interventions for at-risk students. <strong>The</strong> state <strong>of</strong><br />

Andhra Pradesh had conducted a pilot to predict drop-outs based<br />

on past student scores <strong>and</strong> student backgrounds. In one <strong>of</strong> my<br />

previous roles, I had worked towards predicting students’ scores at<br />

Singapore schools, based on historic performance, extra-curricular<br />

activities <strong>and</strong> socio-economic data.<br />

AI <strong>and</strong> allied digital technologies can also be used to mark<br />

students’ answer sheets (including h<strong>and</strong>-written ones), <strong>the</strong>reby<br />

facilitating noting each student’s strengths <strong>and</strong> area <strong>of</strong><br />

improvement. This will save enormous time for teachers by<br />

allowing <strong>the</strong>m to document detailed personalised feedback for<br />

each student.<br />

5. What is <strong>the</strong> role <strong>of</strong> data <strong>and</strong> artificial intelligence<br />

in India’s inclusive growth?<br />

In cases where teachers are not adequately skilled, AI-based<br />

technologies can help <strong>the</strong>m identify areas <strong>of</strong> improvement <strong>and</strong> gain<br />

new skills. AI can be a major tool in ensuring inclusive education<br />

if it is used to identify students who need special teacher<br />

intervention.<br />

7. What are some <strong>of</strong> <strong>the</strong> limitations / challenges that<br />

you foresee with regard to <strong>the</strong> implementation?<br />

<strong>The</strong>se digital technologies would require personalised digital tools<br />

such as computers/laptops/tablets, along with stable internet<br />

connectivity. As mentioned previously, digital access would remain<br />

a major challenge as many rural areas do not yet have access to<br />

quality internet connection or personalised digital devices. Great<br />

progress is being made by <strong>the</strong> government to bridge <strong>the</strong> learning<br />

gaps <strong>and</strong> reap <strong>the</strong> benefits <strong>of</strong> digital technologies.<br />

ISSUE • 9<br />

4 1

11 catalytic sessions with over 50 speakers<br />

THEME 01<br />

THEME 02<br />

THEME 03<br />

THEME 04<br />

THEME 05<br />

Key Updates<br />

at TECH 2018<br />

Transformative<br />

Gaming<br />

<strong>and</strong> Digital<br />

Pedagogies for<br />

Beyond Four Walls<br />

<strong>of</strong> <strong>The</strong> Classroom<br />

Transformative<br />

Gaming<br />

<strong>and</strong> Digital<br />

Pedagogies for<br />

<strong>Artificial</strong><br />

<strong>Intelligence</strong> <strong>and</strong><br />

<strong>the</strong> <strong>Future</strong> <strong>of</strong><br />

<strong>Education</strong><br />

<strong>The</strong> Institutional<br />

Framework for<br />

Application<br />

<strong>of</strong> Digital<br />

SEL<br />

STEM+<br />

Technologies<br />

THEME<br />

in <strong>Education</strong>:<br />

Towards<br />

TECH 2018, UNESCO MGIEP’s international conference, organised with <strong>the</strong> State Government <strong>of</strong> Andhra Pradesh, showcased <strong>the</strong> role<br />

<strong>of</strong> games <strong>and</strong> digital learning in enabling a shift from “transmissive pedagogies” to “transformative pedagogies” to create peaceful<br />

<strong>and</strong> sustainable societies. A key output <strong>of</strong> <strong>the</strong> conference was <strong>the</strong> adoption <strong>of</strong> <strong>the</strong> Vizag Declaration on Guidelines for Digital Learning<br />

as well as <strong>the</strong> signing <strong>of</strong> several Memor<strong>and</strong>ums <strong>of</strong> Underst<strong>and</strong>ing (MOUs) <strong>and</strong> launch announcements.<br />

Surveillance or<br />

Collaborative<br />

<strong>Intelligence</strong>?<br />

mgiep.unesco.org/tech2018<br />

DATES | November 15 – 17, 2018<br />

VENUE | Novotel, Vizag, Andhra Pradesh, India<br />


Jessica Lindl – Global<br />

Head <strong>of</strong> <strong>Education</strong> at Unity<br />

Technologies.<br />

Dan Shefet – Lawyer at<br />

<strong>the</strong> Paris Court <strong>of</strong> Appeal<br />

(France).<br />

Gregoire Borst –<br />

Pr<strong>of</strong>essor <strong>of</strong> Developmental<br />

Psychology <strong>and</strong> Cognitive<br />

Neuroscience <strong>of</strong> <strong>Education</strong>,<br />

LaPsyDE.<br />

Peter Vesterbacka –<br />

Co-founder at Lightneer,<br />

Fun Academy. Formerly<br />

Mighty Eagle at Angry<br />

Birds.<br />

Harri Ketamo – Founder<br />

<strong>and</strong> Chairman, Headai<br />

50+ BREAKOUT<br />


TECH TALK – TECH Talk<br />

by Anantha Duraiappah,<br />

Director, UNESCO MGIEP<br />


PANEL - <strong>Artificial</strong><br />

<strong>Intelligence</strong> in<br />

<strong>Education</strong>: Who<br />

Owns <strong>and</strong> Who<br />

Manages<br />

4 3


Can <strong>Artificial</strong> <strong>Intelligence</strong><br />

Help Us Achieve Universal<br />

Quality <strong>Education</strong>?<br />


James Hodson is a researcher, entrepreneur, <strong>and</strong> social activist. His work spans Machine Learning,<br />

Economics, Sustainable Global Development, Sociology, <strong>and</strong> Philosophy. James currently serves<br />

as CEO <strong>of</strong> <strong>the</strong> AI for Good Foundation; <strong>and</strong> as Senior Researcher at <strong>the</strong> <strong>Artificial</strong> <strong>Intelligence</strong><br />

Department <strong>of</strong> <strong>the</strong> Jozef Stefan Institute, Slovenia. Through <strong>the</strong>se roles, James seeks to effect<br />

lasting change in our economies, societies, <strong>and</strong> scientific underst<strong>and</strong>ing.<br />

Improving Social<br />

Accountability <strong>of</strong><br />

Government Purchases<br />

Through Financial<br />

Statements Data Mining<br />

In recent work from Brazil’s Office <strong>of</strong><br />

<strong>the</strong> Comptroller General, researcher<br />

Rommel Carvalho found that even a<br />

simple implementation <strong>of</strong> an analytical<br />

platform for government transactionlevel<br />

spending could engender more<br />

responsible fiscal behaviour from<br />

<strong>of</strong>ficials. <strong>The</strong> project automatically<br />

identified instances <strong>of</strong> spending<br />

anomalies in procurement systems<br />

(e.g., hundreds to thous<strong>and</strong>s <strong>of</strong> dollars<br />

for a can <strong>of</strong> Coca Cola), <strong>and</strong> allowed<br />

overseers to identify <strong>and</strong> incentivise good<br />

behaviour.<br />

OECD data shows quite clearly a relationship between increased<br />

per capita expenditure on education, <strong>and</strong> increased per capita<br />

GDP. Intuitively, more education is linked to stronger economies,<br />

stronger institutions, <strong>and</strong> stronger societies. However, spending<br />

more doesn’t necessarily translate directly into better education<br />

outcomes, <strong>and</strong> countries with higher perceptions <strong>of</strong> corruption<br />

tend to fare worse, since <strong>the</strong> money that is available is not used<br />

efficiently. Corruption, transparency, <strong>and</strong> accountability are areas<br />

where <strong>Artificial</strong> <strong>Intelligence</strong> can have an important role, ensuring<br />

capital flows to <strong>the</strong> right places <strong>and</strong> is able to impact <strong>the</strong> lives <strong>of</strong><br />

<strong>the</strong> end users--<strong>the</strong> students.<br />

Being inclusive doesn’t just mean solving corrupt<br />

practices in terms <strong>of</strong> monetary flows or systemic<br />

biases. It also means tailoring our education systems to<br />

individual, emotional, cultural, <strong>and</strong> societal needs. <strong>The</strong>se<br />

types <strong>of</strong> adaptation require data <strong>and</strong> knowledge about individuals-<br />

-data that must be protected <strong>and</strong> used only with <strong>the</strong> underst<strong>and</strong>ing<br />

<strong>and</strong> consent <strong>of</strong> <strong>the</strong> humans involved. Privacy, a key factor here, is<br />

yet ano<strong>the</strong>r emerging area <strong>of</strong> research in <strong>Artificial</strong> <strong>Intelligence</strong>--to<br />

ensure we bring <strong>the</strong> benefits <strong>of</strong> <strong>the</strong> data revolution to everyone,<br />

without jeopardising trust.<br />


T<br />

here is hope <strong>and</strong> expectation that <strong>the</strong><br />

maturing <strong>of</strong> <strong>Artificial</strong> <strong>Intelligence</strong> (AI)<br />

techniques over <strong>the</strong> past years could lead to<br />

significant improvements across <strong>the</strong> social<br />

spectrum. From improving global health <strong>and</strong><br />

well-being to alleviating food security concerns, preventing crime,<br />

<strong>and</strong> addressing corruption, AI has been heralded as a universal<br />

solution. At <strong>the</strong> AI for Good Foundation, we measure <strong>the</strong> potential<br />

for <strong>Artificial</strong> <strong>Intelligence</strong> to have an impact to achieve <strong>the</strong> Sustainable<br />

Development Goals. One <strong>of</strong> <strong>the</strong>se goals is to ensure quality<br />

<strong>and</strong> inclusive education for all, <strong>and</strong> promote lifelong learning.<br />

However, learning has been one <strong>of</strong> <strong>the</strong> areas least impacted by<br />

<strong>the</strong> onslaught <strong>of</strong> data science <strong>and</strong> machine learning--with little to<br />

speak <strong>of</strong> beyond light personalisation <strong>of</strong> Massive Open Online<br />

Courses (MOOCs).<br />

It would be unfair to say that no effort is being dedicated to this<br />

task. Certainly, <strong>the</strong>re are researchers working on virtual<br />

tutors 1 , automated text-book creation 2 , personalised<br />

learning paths 3 <strong>and</strong> more 4 . However little as yet has made it<br />

into mainstream adoption, <strong>and</strong> even less has been shown to have<br />

consistent positive impact on learning outcomes. Classrooms,<br />

training <strong>and</strong> continuous education in developed countries (those<br />

with ample resources to experiment) look more or less as <strong>the</strong>y did<br />

50 years ago, except with access to <strong>the</strong> internet, <strong>and</strong> perhaps some<br />

“educational activities” on computers that do little more than<br />

distract classrooms <strong>and</strong> disrupt social learning.<br />

What can <strong>and</strong> should we do to bring <strong>the</strong> benefits <strong>of</strong> big<br />

data <strong>and</strong> <strong>Artificial</strong> <strong>Intelligence</strong> to improve quality <strong>of</strong><br />

learning?<br />

More Open Data, More Open <strong>Education</strong><br />

Resources<br />

In online settings, massive amounts <strong>of</strong> data are being collected<br />

regarding learner choices, behaviours, <strong>and</strong> outcomes. Some<br />

<strong>of</strong> <strong>the</strong>se data are being used to make recommendations, or<br />

to improve courses in future. <strong>The</strong>se data need to be properly<br />

harnessed to drive our underst<strong>and</strong>ing <strong>of</strong> how individual students<br />

learn best, <strong>and</strong> to give <strong>the</strong>m <strong>the</strong> guidance <strong>and</strong> resources <strong>the</strong>y need<br />

on a case by case, self-serve basis. Learning settings need to escape<br />

<strong>the</strong> classroom into <strong>the</strong> real world--allowing people to continue<br />

learning along <strong>the</strong>ir personalised paths to <strong>the</strong>ir learning goals at<br />

any time.<br />

Intelligent, Supportive Learning<br />

Environments<br />

In classroom settings, teachers <strong>of</strong>ten do not have <strong>the</strong><br />

time for highly personalised 1-1 tutoring <strong>and</strong> feedback.<br />

Instead, work is set to be reviewed at a later time, introducing a lag<br />

between effort <strong>and</strong> feedback. This separation dilutes <strong>the</strong> learning<br />

experience, <strong>and</strong> makes it harder for teachers to tailor content to<br />

<strong>the</strong>ir students. Classroom support “AI Tutor” systems have been<br />

available since <strong>the</strong> 1970s’, <strong>and</strong> yet today’s most sophisticated<br />

learning technologies (e.g., Rosetta Stone’s language learning<br />

s<strong>of</strong>tware) use only very basic machine learning (for example, to<br />

match speech patterns) to help students succeed. Fifty years on,<br />

learning environments should be immersive, interactive, <strong>and</strong><br />

virtual tutors should be <strong>the</strong>re for students whenever a teacher<br />

cannot be, promoting social, emotional <strong>and</strong> conceptual learning.<br />

Promote Economic Opportunity<br />

Underst<strong>and</strong>ing <strong>the</strong> relationship between learning, skills, <strong>and</strong><br />

employment is critical for <strong>the</strong> future strength <strong>of</strong> our economy.<br />

<strong>Artificial</strong> <strong>Intelligence</strong> can help us to identify trends from<br />

local to global scales, <strong>and</strong> prepare <strong>the</strong> workforce <strong>of</strong><br />

tomorrow today. Not only can a detailed view <strong>of</strong> <strong>the</strong> current<br />

mosaic <strong>of</strong> supply <strong>and</strong> dem<strong>and</strong> <strong>of</strong> abilities help us to ensure<br />

consistent high quality employment, but also to appropriately<br />

incentivise <strong>the</strong> learning <strong>of</strong> skills that will create economic<br />

opportunity in future. Labour economists <strong>and</strong> AI researchers can<br />

work h<strong>and</strong> in h<strong>and</strong> to ensure appropriate models are built <strong>and</strong><br />

mechanisms derived to extract maximum social surplus.<br />

Even <strong>the</strong> best systems are useless if <strong>the</strong>y don’t reach <strong>the</strong><br />

people who need <strong>the</strong>m most. Can AI help us to improve<br />

global access to learning?<br />

Eradicate Corruption<br />

<strong>Education</strong> is <strong>the</strong> cornerstone <strong>of</strong> human prosperity <strong>and</strong><br />

flourishing. <strong>Education</strong> brings underst<strong>and</strong>ing, compassion, <strong>and</strong><br />

<strong>the</strong> means to make a positive difference in <strong>the</strong> world. Now is<br />

<strong>the</strong> time to bring <strong>the</strong> benefits <strong>of</strong> technology, data <strong>and</strong><br />

<strong>Artificial</strong> <strong>Intelligence</strong> to learning at all levels, in all<br />

environments <strong>and</strong> for all people. With focused efforts today,<br />

we can break down <strong>the</strong> barriers, inequities <strong>and</strong> misunderst<strong>and</strong>ings<br />

that prevent us from achieving our full potential. <strong>Artificial</strong><br />

<strong>Intelligence</strong> is not a one-stop solution, but its presence can<br />

allow us to place renewed emphasis on <strong>the</strong> identification <strong>and</strong><br />

implementation <strong>of</strong> new paradigms for learning <strong>and</strong> development.<br />

Let’s use this as an opportunity to take big steps forward.<br />

About <strong>the</strong> AI for Good Foundation<br />

<strong>The</strong> AI for Good Foundation is a US 501(c)3 Public Charity <strong>and</strong><br />

global non-pr<strong>of</strong>it that fosters activities to maximise <strong>the</strong> benefit<br />

<strong>of</strong> AI technologies for social good through <strong>the</strong> lens <strong>of</strong> global<br />

sustainable development. Through a blend <strong>of</strong> foundational<br />

research, outreach, education <strong>and</strong> policy engagement our current<br />

projects are helping to advance <strong>the</strong> global sustainable development<br />

agenda — a set <strong>of</strong> goals adopted by 150 countries to end poverty,<br />

protect <strong>the</strong> planet, <strong>and</strong> ensure prosperity for all. Find out more<br />

about <strong>the</strong> AI for Good Foundation <strong>and</strong> how you can help by<br />

visiting https://www.ai4good.org/.<br />

ISSUE • 9<br />

1.<br />

Johnson, W.L. & Lester, J.C. “Face-to-Face Interaction with Pedagogical Agents, Twenty Years Later”,<br />

Int J Artif Intell Educ (2016) 26: 25. https://doi.org/10.1007/s40593-015-0065-9<br />

2.<br />

Singh, Rohit, Sumit Gulwani, <strong>and</strong> Sriram K. Rajamani. “Automatically Generating Algebra Problems.” AAAI. 2012.<br />

3.<br />

Chen, Chih-Ming. “Intelligent web-based learning system with personalised learning path guidance.” Computers & <strong>Education</strong> 51.2 (2008): 787-814.<br />

4.<br />

See, for example, “<strong>Artificial</strong> <strong>Intelligence</strong> in <strong>Education</strong>”, C.-K. Looi et. al. (Eds.), IOS Press (2005). 4 5




“AI for Good Foundation.” AI for Good Foundation, ai4good.org/.<br />

Carvalho, Rommel N., et al. “Probabilistic Ontology <strong>and</strong> Knowledge Fusion for Procurement<br />

Fraud Detection in Brazil.” Uncertainty Reasoning for <strong>the</strong> Semantic Web II Lecture Notes in<br />

Computer Science, 2013, pp. 19–40., doi:10.1007/978-3-642-35975-0_2.<br />

Carvalho, Rommel, et al. “Using Clustering <strong>and</strong> Text Mining to Create a Reference Price<br />

Database.” Learning <strong>and</strong> Nonlinear Models, vol. 12, no. 1, 2014, pp. 38–52., doi:10.21528/<br />

lnlm-vol12-no1-art3.<br />

Chen, Chih-Ming. “Intelligent Web-Based Learning System with Personalised Learning Path<br />

Guidance.” Computers &amp; <strong>Education</strong>, vol. 51, no. 2, 2008, pp. 787–814., doi:10.1016/j.<br />

compedu.2007.08.004.<br />

Fairfield, John. “Speech Comparison in <strong>the</strong> Rosetta Stone” Proceedings <strong>of</strong> a Symposium on<br />

Computer Mediated Language Assessment <strong>and</strong> Evaluation in Natural Language Processing -<br />

ASSESSEVALNLP ‘99, 1999, doi:10.3115/1598834.1598837.<br />

Farzindar, Atefeh, <strong>and</strong> Kešelj Vlado. Advances in <strong>Artificial</strong> <strong>Intelligence</strong>: 23rd Canadian<br />

Conference on <strong>Artificial</strong> <strong>Intelligence</strong>, Canadian AI 2010, Ottawa, Canada, May 31 - June 2,<br />

2010: Proceedings. Springer, 2010.<br />

“Home .:. Sustainable Development Knowledge Platform.” United Nations, United Nations,<br />

sustainabledevelopment.un.org/.<br />

Hoppe, Ulrich. <strong>Artificial</strong> <strong>Intelligence</strong> in <strong>Education</strong>: Shaping <strong>the</strong> <strong>Future</strong> <strong>of</strong> Learning through<br />

Intelligent Technologies. IOS Press, 2003.<br />

Johnson, W. Lewis, <strong>and</strong> James C. Lester. “Face-to-Face Interaction with Pedagogical Agents,<br />

Twenty Years Later.” International Journal <strong>of</strong> <strong>Artificial</strong> <strong>Intelligence</strong> in <strong>Education</strong>, vol. 26, no.<br />

1, 2015, pp. 25–36., doi:10.1007/s40593-015-0065-9.<br />

Lee, Fong-Lok. “Web Algebra Tutor: A Web-Based AI-Enabled Interactive Learning<br />

Environment in Algebra.” <strong>The</strong> International Journal <strong>of</strong> Learning: Annual Review, vol. 8, no.<br />

1, 2004, doi:10.18848/1447-9494/cgp/v08/47970.<br />

Li, Tao, <strong>and</strong> Samuel Sambasivam. “Automatically Generating Questions in Multiple<br />

Variables for Intelligent Tutoring.” Proceedings <strong>of</strong> <strong>the</strong> 2005 InSITE Conference, 2005,<br />

doi:10.28945/2890.<br />

Looi, Chee-Kit. <strong>Artificial</strong> <strong>Intelligence</strong> in <strong>Education</strong>: Supporting Learning through Intelligent<br />

<strong>and</strong> Socially Informed Technology. IOS, 2005.<br />

Rosé, Carolyn Penstein. <strong>Artificial</strong> <strong>Intelligence</strong> in <strong>Education</strong>: 19th International Conference,<br />

AIED 2018, London, UK, June 27-30, 2018: ProceedingsnPart 1. Springer, 2018.<br />

“THE VIRTUAL TUTOR - Development <strong>of</strong> a Virtual Tutor within a Simulated Working<br />

Environment.” Proceedings <strong>of</strong> <strong>the</strong> 4th International Conference on Computer Supported<br />

<strong>Education</strong>, 2012, doi:10.5220/0003966601910194.<br />

Weischedel, Ralph M., et al. “An <strong>Artificial</strong> <strong>Intelligence</strong> Approach to Language Instruction.”<br />

<strong>Artificial</strong> <strong>Intelligence</strong>, vol. 10, no. 3, 1978, pp. 225–240., doi:10.1016/s0004-3702(78)80015-<br />

0.<br />

Wenger, Etienne. <strong>Artificial</strong> <strong>Intelligence</strong> <strong>and</strong> Tutoring Systems: Computational <strong>and</strong> Cognitive<br />

Approaches to <strong>the</strong> Communication <strong>of</strong> Knowledge. Morgan Kaufmann Publishers, 1987.<br />

<strong>Artificial</strong> <strong>Intelligence</strong><br />

for <strong>Education</strong><br />


Premise <strong>and</strong> Problem<br />

Free <strong>and</strong> Compulsory <strong>Education</strong> for children in <strong>the</strong> age group <strong>of</strong><br />

6-14 is a Fundamental Right, enacted by <strong>the</strong> Parliament <strong>of</strong> India<br />

[1]. It acknowledges, in spirit, that access to education is necessary<br />

to realising an individual’s inner potential. However, it is quite a<br />

feat to achieve even this relatively modest goal in countries such<br />

as India, where <strong>the</strong> varied demographic l<strong>and</strong>scape renders “one<br />

size fits all solutions” ineffective [2]. In addition, <strong>the</strong> number <strong>of</strong><br />

<strong>the</strong> aspirants, <strong>and</strong> <strong>the</strong> unfavourable skew on <strong>the</strong> supply side <strong>of</strong> <strong>the</strong><br />

quality resources, makes it even harder. A larger supporting<br />

ecosystem is needed, wherein an individual is given<br />

attention based on his/her learning capabilities <strong>and</strong> <strong>the</strong><br />

changing world order. Adequate quality learning resources at<br />

all levels (including physical, emotional, individual <strong>and</strong> societal)<br />

are also required. Can technology help?<br />

Prior<br />

For a long time, EdTech remained an academic activity, where<br />

both models <strong>of</strong> learning materials <strong>and</strong> methods, <strong>and</strong> <strong>the</strong> means<br />

<strong>of</strong> instructional delivery, such as those found in intelligent tutoring<br />

systems, were largely experimental <strong>and</strong> were without much<br />

penetration, despite strong <strong>the</strong>oretical foundations [3-4]. However,<br />

as <strong>the</strong> internet has become both accessible <strong>and</strong> affordable,<br />

Soma S Dhavala is an Innovator, Entrepreneur <strong>and</strong> AI Architect, operating at <strong>the</strong> interface <strong>of</strong><br />

Statistics, Machine Learning, Computing <strong>and</strong> <strong>the</strong> Internet <strong>of</strong> Things. He is a co-founder <strong>of</strong><br />

VitalTicks, a digital healthcare startup, - where he is leading product design, architecture <strong>and</strong><br />

solutioning. Soma consults in <strong>the</strong> EdTech space, architecting AI solutions <strong>and</strong> platforms, <strong>and</strong><br />

designs <strong>and</strong> reviews curriculum in AI <strong>and</strong> Deep Learning.<br />

Soma obtained his Ph.D (TAMU) in Statistics (where he worked on applying Bayesian<br />

Nonparametrics to systems biology) in 2010, Master’s (IIT-M) in 2000 <strong>and</strong> B.E (SRKREC, Andhra<br />

Pradesh) in 1997. He also worked in a post-doctoral role in <strong>the</strong> area <strong>of</strong> Dynamical Systems at<br />

TAMU for almost a year. He has over 10 publications <strong>and</strong> multiple patents to his credit.<br />

MOOCs <strong>and</strong> Flipped Classrooms have become successful, <strong>and</strong><br />

access to quality content is no longer a barrier, at least in terms<br />

<strong>of</strong> higher education. A person in a remote village in India can<br />

enroll in a course taught <strong>and</strong> conducted by faculty at Stanford,<br />

which would have been unthinkable a few years ago. Despite <strong>the</strong><br />

credibility <strong>of</strong> <strong>the</strong> <strong>of</strong>fering <strong>and</strong> <strong>the</strong> relatively privileged learning<br />

groups, such platforms suffer heavy dropout rates, largely because<br />

ei<strong>the</strong>r <strong>the</strong> intrinsic needs or <strong>the</strong> necessary prerequisites, or both,<br />

have not been met. As per <strong>the</strong> ASER 2017 report, unaddressed<br />

learning gaps typically carry forward, making <strong>the</strong> learner even<br />

more vulnerable to failure in future [6]. Given a deeper mobile<br />

penetration, <strong>and</strong> cheaper data access plans, can datahungry<br />

AI solutions deliver personalised education, <strong>and</strong><br />

rival <strong>the</strong> breakthroughs happening in o<strong>the</strong>r disciplines<br />

such as Computer Vision <strong>and</strong> Natural Language<br />

Underst<strong>and</strong>ing (NLU)?<br />

Parts<br />

AAI-powered or an o<strong>the</strong>rwise learner-centric EdTech<br />

<strong>of</strong>fering needs to have, at <strong>the</strong> least, 1) presentable content <strong>and</strong><br />

representation <strong>of</strong> its constituents 2) an assessment framework 3)<br />

a taxonomy <strong>of</strong> <strong>the</strong> concepts <strong>and</strong> lesson plans to chart learning<br />

journeys. Outside <strong>the</strong> realm <strong>of</strong> <strong>the</strong>se formal pedagogic elements<br />


ISSUE • 9<br />

4 7


In order for AI-powered solutions to be<br />

contextualised, commoditised, <strong>and</strong><br />

available to wider, diverse audience, a rich<br />

set <strong>of</strong> tooling, st<strong>and</strong>ards, <strong>and</strong> best practices<br />

are required.<br />



lie <strong>the</strong> important but <strong>of</strong>ten overlooked elements such as 4) user<br />

experience, 5) behavioural <strong>and</strong> social contexts, <strong>and</strong> 6) benevolent<br />

marketing <strong>and</strong> br<strong>and</strong>ing. Without <strong>the</strong>se conditions, chances <strong>of</strong><br />

an EdTech solution being a successful product are ra<strong>the</strong>r slim.<br />

None<strong>the</strong>less, <strong>the</strong>re are efforts to push <strong>the</strong> envelope, <strong>and</strong> harness<br />

<strong>the</strong> seemingly ripe conditions for trying out AI powered solutions<br />

[7-10]. Any particular solution makes certain assumptions<br />

about those elements in <strong>the</strong> composition. Consequentially, <strong>the</strong><br />

relevance <strong>and</strong> effectiveness vary for any learner. In order for AIpowered<br />

solutions to be contextualised, commoditised,<br />

<strong>and</strong> available to wider, diverse audience, a rich set <strong>of</strong><br />

tooling, st<strong>and</strong>ards, <strong>and</strong> best practices are required.<br />

Product Building<br />

AI <strong>and</strong> Machine Learning are still evolving. Unlike physical<br />

products that can be perceived directly, such as an electric<br />

fan, AI products are most likely be intangible. For example,<br />

a manufacturer can provide a five-year warranty on <strong>the</strong> fan’s<br />

motor. A user can exercise <strong>the</strong> warranty when <strong>the</strong> motor burns<br />

out. Contrast that with an AI-powered content recommendation<br />

solution – what can be guaranteed or certified about it? This<br />

opens up a p<strong>and</strong>ora’s box. Is <strong>the</strong>re a way <strong>of</strong> building AI-powered<br />

products so that <strong>the</strong>y conform to st<strong>and</strong>ards <strong>and</strong> scrutiny by<br />

third-parties? Are <strong>the</strong> results reproducible, auditable, scalable <strong>and</strong><br />

customisable, among many o<strong>the</strong>r dimensions? <strong>The</strong> AI industry<br />

is responding to <strong>the</strong>se challenges with conceptual<br />

frameworks [12], open st<strong>and</strong>ards <strong>and</strong> platforms [13-15],<br />

<strong>and</strong> AI education resources [16]. Ano<strong>the</strong>r set <strong>of</strong> orthogonal<br />

concerns has far-reaching implications from a pedagogical<br />

<strong>and</strong> ethical st<strong>and</strong>point. At a fundamental level, can <strong>the</strong> results<br />

be explained to <strong>the</strong> learner? Can <strong>the</strong> Right to Explanation be<br />

enforced, which could be enacted as a law in <strong>the</strong> future? Should<br />

AI-driven EdTech be regulated to safeguard learner interests? For<br />

example, a popular commercially available learning application is<br />

not in <strong>the</strong> purview <strong>of</strong> regulation.<br />

Conclusion<br />

<strong>The</strong> potential <strong>of</strong> personalised quality education, regardless <strong>of</strong><br />

geographical, socio-cultural <strong>and</strong> o<strong>the</strong>r affiliations etc.., by means<br />

<strong>of</strong> AI <strong>and</strong> Deep Learning, is recognised [19-22]. However, a<br />

sustained, coordinated effort to address <strong>the</strong>m holistically, from<br />

pedagogical to experiential to business to ethics <strong>and</strong> governance<br />

angles, is required. Concerted efforts are needed to develop<br />

a) data-first platforms b) shared NLU <strong>and</strong> AI infrastructure <strong>and</strong><br />

c) agile processes to develop AI products. <strong>The</strong>se products should<br />

eventually be used to create customised, relevant, engaging,<br />

explainable <strong>and</strong> governable solutions for every learner.<br />

[1] Right To <strong>Education</strong> Act, 2009<br />

[2] Still too many children out <strong>of</strong> school, <strong>The</strong> Hindu, September 4, 2018<br />

[3] Nikos, M., Hendrik, D., Katrien, V., <strong>and</strong> Erik, Duval., (2014), Recommender<br />

Systems for Learning, Springer-Verlag NewYork, 2014.<br />

[4] Advances in Intelligent Tutoring Systems, Eds. Editors: Nkambou, Roger,<br />

Mizoguchi, Riichiro, Bourdeau, Jacqueline (Eds.), Springer-Verlag Berlin<br />

Heidelberg, 2010.<br />

[5] Jordan, K. (2015). Massive open online course completion rates revisited:<br />

Assessment, length <strong>and</strong> attrition. <strong>The</strong> International Review <strong>of</strong> Research in Open<br />

<strong>and</strong> Distributed Learning<br />

[6] ASER 2017: Beyond Basics<br />

[7] ALEKS<br />

[8] ENLEARN<br />

[9] Sana Labs: AI for <strong>Education</strong><br />

[10] expi.com<br />

[11] ReedEx: An Intelligent Reading Assistant<br />

[12] Fifteen Comm<strong>and</strong>ments <strong>and</strong> an Axiomatic Framework for building Machine<br />

Learning Products<br />

[13] openAI: Discovering <strong>and</strong> enacting <strong>the</strong> path to safe artificial general intelligence<br />

[14] AllenAI: AI for <strong>the</strong> common good<br />

[15] Wadhvani.ai: AI for social good<br />

[16] 500k.ai: 500 Thous<strong>and</strong> AI Programmers Initiative<br />

[17] Byju’s personalization engine<br />

[18] <strong>The</strong> EU General Data Protection Regulation<br />

[19] Dhar, V., Nilekani, N., S. Maruwada, Pappu, N. (2016), Big Data as an Enabler,<br />

Journal <strong>of</strong> Big Data, Vol. 4, No. 3, pp. 137-140<br />

[20] Kurshan, B, <strong>The</strong> <strong>Future</strong> <strong>of</strong> <strong>Artificial</strong> <strong>Intelligence</strong> in <strong>Education</strong>, in Forbes,<br />

accessed on 6th September, 2018.<br />

[21] China State Council’s “A Next Generation AI Development Plan”,<br />

[22] NITI Ayog’s discussion paper on “National Strategy for <strong>Artificial</strong> <strong>Intelligence</strong><br />

document”, 2018<br />

All links were accessed on 6th September, 2018<br />

ISSUE • 9<br />

4 9



Dr. Anantha Duraiappah elected as a Fellow<br />

to <strong>The</strong> World Academy <strong>of</strong> Sciences<br />

May, 2018<br />

Dr. Anantha K. Duraiappah, Director, UNESCO Mahatma<br />

G<strong>and</strong>hi Institute <strong>of</strong> <strong>Education</strong> for Peace <strong>and</strong> Sustainable<br />

Development (MGIEP) was recently elected as a Fellow to<br />

<strong>The</strong> World Academy <strong>of</strong> Sciences (TWAS), a pioneering,<br />

merit-based academy focused on <strong>the</strong> advancement <strong>of</strong> science<br />

in developing countries. Dr. Duraiappah was recognised <strong>and</strong><br />

elected to TWAS for his contribution to <strong>the</strong> progression <strong>of</strong><br />

<strong>the</strong> field <strong>of</strong> sciences in developing nations.Speaking on his<br />

election, Dr. Duraiappah expressed: “It is an honour to be<br />

a part <strong>of</strong> this prestigious body <strong>of</strong> scientists who have carried<br />

out some ground-breaking work in science. I look forward<br />

to working alongside this body towards advancement <strong>of</strong><br />

sciences – specifically in <strong>the</strong> field <strong>of</strong> learning sciences.”<br />

UNESCO MGIEP conducts a Plenary Session<br />

at <strong>the</strong> <strong>the</strong> 11 th Biennial Comparative<br />

<strong>Education</strong> Society <strong>of</strong> Asia (CESA) Conference<br />

May 11-12, 2018, Siem Reap, Cambodia<br />

<strong>The</strong> Comparative <strong>Education</strong> Society <strong>of</strong> Asia (CESA) is an<br />

association <strong>of</strong> educational scholars, working to promote<br />

comparative education as a scholarly field across Asia<br />

<strong>and</strong> to streng<strong>the</strong>n Asian voices in global debates about<br />

education. <strong>The</strong> 11th CESA Biennial Conference was held<br />

on 11–12 May 2018 in Siem Reap, Cambodia. <strong>The</strong> <strong>the</strong>me<br />

for 2018 was ‘<strong>Education</strong> <strong>and</strong> Social Progress: Insights from<br />

Comparative Perspectives’. On May 12, a plenary panel<br />

discussion was dedicated to UNESCO MGIEP’s seminal<br />

report Rethinking Schooling: <strong>The</strong> State <strong>of</strong> <strong>Education</strong> for<br />

Peace, Sustainable Development <strong>and</strong> Global Citizenship in<br />

Asia. Dr Yoko Mochizuki, Head <strong>of</strong> Rethinking Curricula<br />

Programme <strong>of</strong> UNESCO MGIEP, presented <strong>the</strong> key<br />

findings.<br />


<strong>The</strong> Government <strong>of</strong> Delhi organises a<br />

workshop for Special Educators in Screening<br />

Different Learners<br />

May 10, 2018, New Delhi, India<br />

In its mission to achieve inclusive <strong>and</strong> quality education for<br />

all (SDG 4.7), UNESCO MGIEP conducted a capacitybuilding<br />

workshop to enhance <strong>the</strong> skills <strong>of</strong> special educators<br />

in identifying <strong>and</strong> addressing students who learn differently<br />

as part <strong>of</strong> Government <strong>of</strong> Delhi’s project SMILE for its<br />

public schools. It was held on May 10, at Department <strong>of</strong><br />

<strong>Education</strong>’s premises <strong>and</strong> attended by 60 psychologists<br />

from <strong>the</strong> National Capital Region. UNESCO MGIEP’s<br />

Difference Learning Project aims to improve functional<br />

literacy rate by enabling teachers, special educators,<br />

psychological experts <strong>and</strong> parents to identify different<br />

learners, provide diagnosis <strong>and</strong> cater to <strong>the</strong>ir different<br />

needs in <strong>the</strong> classroom. Dr N<strong>and</strong>ini C. Singh, Head <strong>of</strong><br />

<strong>the</strong> Difference Learning Project at UNESCO MGIEP,<br />

conducted <strong>the</strong> workshop in collaboration with Ms Geet<br />

Oberoi, Founder <strong>of</strong> Orkids, an organisation working in<br />

<strong>the</strong> field <strong>of</strong> assessment, remediation <strong>and</strong> teacher training<br />

<strong>of</strong> educators.were discussed. <strong>The</strong> meeting also provided<br />

<strong>the</strong> participants with <strong>the</strong> opportunity to underst<strong>and</strong> <strong>the</strong><br />

work <strong>of</strong> <strong>the</strong>ir fellow partners for new collaborations <strong>and</strong><br />

partnerships to emerge.<br />

Textbook Embedding Project going strong<br />

in Sikkim with <strong>the</strong> Second workshop for<br />

Textbook Authors<br />

May 20 – 22, 2018, Gangtok, India<br />

<strong>The</strong> project to embed concepts <strong>of</strong> peace <strong>and</strong> sustainable<br />

development in textbooks <strong>of</strong> Sikkim is progressing steadily<br />

with <strong>the</strong> second workshop for textbook authors organised in<br />

Gangtok from May 20-22, 2018. <strong>The</strong> embedding project is<br />

conceptualised as a continuous <strong>and</strong> comprehensive capacity<br />

building <strong>of</strong> textbook authors. <strong>The</strong> second workshop<br />

marked a beginning with textbook authors starting to create<br />

early drafts <strong>of</strong> examples <strong>of</strong> chapters, which went through<br />

participatory feedbacks as well as feedback from facilitators.<br />

ISSUE • 9<br />

5 1



YESPeace Partners’ Meeting- India<br />

May 22-24, 2018, Gurugram, India<br />

UNESCO MGIEP hosted <strong>the</strong> 2018 Partners Meeting<br />

<strong>of</strong> <strong>the</strong> Youth for <strong>Education</strong>, Sustainability <strong>and</strong> Peace<br />

(YESPeace)Network at TERI Gram in Gurugram, India<br />

from 22-24 May 2018. <strong>The</strong> Partners’ Meeting brought<br />

toge<strong>the</strong>r existing <strong>and</strong> potential partners <strong>of</strong> <strong>the</strong> YESPeace<br />

Network from across South Asia <strong>and</strong> Africa. <strong>The</strong> meeting<br />

served as a platform to develop a common <strong>and</strong> shared<br />

underst<strong>and</strong>ing <strong>of</strong> <strong>the</strong> YESPeace network, discuss <strong>the</strong><br />

potential <strong>of</strong> cross-country <strong>and</strong> intra-country collaboration<br />

between YESPeace partners, identify timelines <strong>and</strong> orient<br />

<strong>the</strong> partners to MGIEP’s Youth Kindness Campaign.<br />

Different ideas <strong>and</strong> strategies were explored <strong>and</strong> plans<br />

for <strong>the</strong> growth <strong>and</strong> expansion <strong>of</strong> <strong>the</strong> network were<br />

discussed. <strong>The</strong> meeting also provided participants with <strong>the</strong><br />

opportunity to underst<strong>and</strong> <strong>the</strong> work <strong>of</strong> <strong>the</strong>ir fellow partners<br />

for new collaborations <strong>and</strong> partnerships to emerge.<br />

How should we prevent violent extremism?<br />

<strong>The</strong> youth may have <strong>the</strong> answer!<br />

May 25, 2018, New Delhi, India<br />

It may have been early on a Friday morning, but<br />

<strong>the</strong> 100 plus youth who attended <strong>the</strong> launch <strong>of</strong><br />

#YouthWagingPeace—Action Guidelines for <strong>the</strong> Prevention<br />

<strong>of</strong> Violent Extremism (PVE) were raring to go with ideas,<br />

questions <strong>and</strong> tangible action points for PVE. <strong>The</strong> launch<br />

was followed by a riveting panel discussion on Youth<br />

Action Across Borders which saw panellists such as author<br />

<strong>and</strong> youth activist Gurmehar Kaur from India diving<br />

deep into <strong>the</strong>ir own experiences on youth empowerment.<br />

Underscoring <strong>the</strong> need for action-oriented guidelines,<br />

H.E Ms. Harinder Sidhu, High Commissioner to India,<br />

Australian High Commission in New Delhi said, “<strong>The</strong><br />

#YouthWagingPeace Action Guidelines helps individuals<br />

<strong>and</strong> communities to build tools such as critical thinking,<br />

resilience, compassion <strong>and</strong> respect for diversity which <strong>the</strong>y<br />

need to challenge extremist ideologies.”<br />


Three-day workshop for young educators on<br />

prevention <strong>of</strong> violent extremism<br />

May 22-24, 2018, Gurgaon, India<br />

<strong>The</strong> launch event <strong>of</strong> #YouthWagingPeace – Action<br />

Guidelines for Prevention <strong>of</strong> Violent Extremism was<br />

preceded by an intensive 3-day workshop for 25 young<br />

educators from across India, including three Rohingya<br />

youth working for <strong>the</strong> Rohingya refugee community in<br />

India.<br />

Facilitated by Ms Carolyn Nash, Coordinating Lead Author,<br />

#YouthWagingPeace <strong>and</strong> Mr Simon Kuany, Project Officer,<br />

#YouthWagingPeace, <strong>the</strong> 25 young educators were able to<br />

adapt <strong>the</strong> Action Guidelines <strong>and</strong> came up with <strong>the</strong>ir own<br />

context-specific lesson plans for <strong>the</strong>ir respective stakeholders<br />

(teachers, school administrators, families <strong>and</strong> policymakers).<br />

UNESCO MGIEP will continue to work with <strong>the</strong>se young<br />

educators to ensure that <strong>the</strong> lesson plans are implemented<br />

<strong>and</strong> <strong>the</strong> outputs reflected/included in <strong>the</strong> broader<br />

#YouthWagingPeace workbook. This workbook will include<br />

all <strong>the</strong> tested lesson plans developed <strong>and</strong> tried by young<br />

participants <strong>of</strong> #YouthWagingPeace workshops.<br />

Learn to play <strong>and</strong> play to learn – a<br />

Distinguished Lecture hosted by UNESCO<br />

MGIEP on <strong>the</strong> role <strong>of</strong> gaming in education<br />

June 28, 2018, New Delhi, India<br />

Can learning be made fun <strong>and</strong> experiential? In UNESCO<br />

MGIEP’s Distinguished Lecture held at <strong>the</strong> UNESCO<br />

New Delhi, India on Thursday, June 28, 2018, Zigor<br />

Hern<strong>and</strong>orena Juarros, Senior Project Manager – Fun<br />

Learning Department, Ubis<strong>of</strong>t (makers <strong>of</strong> Assassins Creed<br />

Franchise, <strong>The</strong> Division, Far Cry, Watch Dog), discussed<br />

how digital pedagogies such as games can be used as an<br />

effective tool to impart learning. <strong>The</strong> lecture was attended<br />

by over a hundred participants including teachers,<br />

educators, curriculum designers, ed-tech experts, games<br />

designers <strong>and</strong> developers as well as students. In his lecture,<br />

titled ‘Learn to play <strong>and</strong> play to learn, <strong>the</strong> role <strong>of</strong> gaming<br />

in education’, Zigor explained how digital games can be<br />

designed to reward players, acknowledge practice <strong>and</strong><br />

incremental progress, thus making learning more effective<br />

through peer-collaboration <strong>and</strong> a project-based approach.<br />

Zigor also underlined <strong>the</strong> need to develop suitable content<br />

for games to make <strong>the</strong>m more relevant to 21 st century<br />

learners’ reaction from vision, <strong>and</strong> to turn <strong>the</strong> victim into<br />

<strong>the</strong> hero or <strong>the</strong> warrior for waging peace.<br />

ISSUE • 9<br />

5 3



Youth for SDGs Peace Tables: Organised by<br />

St<strong>and</strong>ing Toge<strong>the</strong>r to Enable Peace<br />

June 29 - July 9, 2018, New Delhi, India<br />

Twenty youth leaders from New Delhi <strong>and</strong> o<strong>the</strong>r parts <strong>of</strong><br />

India have created dynamic Community Action Plans for<br />

Art <strong>and</strong> Democracy, digital media, gender, <strong>and</strong> interfaith<br />

dialogue. <strong>The</strong> plans were developed through a series <strong>of</strong><br />

Peace Tables from June to July 2018, organised by St<strong>and</strong>ing<br />

Toge<strong>the</strong>r to Enable Peace (STEP), a prominent Indian<br />

NGO working to build peace <strong>and</strong> a partner in UNESCO<br />

MGIEP’s Youth for <strong>Education</strong>, Sustainability <strong>and</strong> Peace<br />

(YESPeace) Network. Through <strong>the</strong> process <strong>of</strong> <strong>the</strong> Peace<br />

Tables, participants use <strong>the</strong>ir skills to combat <strong>the</strong> problems<br />

<strong>of</strong> incomplete <strong>and</strong> unequal implementation <strong>of</strong> <strong>the</strong> United<br />

Nations Sustainable Development Goals (SDGs) as well<br />

as <strong>the</strong> challenges to peace, prosperity <strong>and</strong> security in <strong>the</strong><br />

context <strong>of</strong> a developing country like India. <strong>The</strong> Peace<br />

Tables successfully enabled <strong>the</strong> young leaders to dwell deep<br />

into <strong>the</strong>ir motivations, to sieve out reaction from vision, <strong>and</strong><br />

to turn <strong>the</strong> victim into <strong>the</strong> hero or <strong>the</strong> warrior for waging<br />

peace.<br />

First ‘Training <strong>the</strong> Trainers Workshop’ on<br />

Cantor’s World<br />

July 24 - 25, 2018, New Delhi, India<br />

Second <strong>Future</strong>s Workshop on ‘<strong>The</strong> State<br />

<strong>of</strong> Social <strong>and</strong> Emotional Learning (SEL)<br />

Assessment’<br />

July 19 - 21, 2018, Paris, France<br />

UNESCO MGIEP organised <strong>the</strong> Second <strong>Future</strong>s’<br />

Workshop, held at La Sorbonne University, Paris between<br />

July 19-21, 2018, bringing toge<strong>the</strong>r an expert group<br />

comprising <strong>of</strong> representatives <strong>of</strong> different disciplines,<br />

including neuroscience, human development psychology,<br />

anthropology, <strong>and</strong> education. <strong>The</strong> <strong>the</strong>mes discussed<br />

were evidence-based practices for SEL implementation<br />

in education, <strong>and</strong> <strong>the</strong> commitment to bring nuanced<br />

cultural contextualisation to <strong>the</strong> implementation <strong>of</strong> SEL<br />

programming in schools.<br />


Compassion Acts, Empathy <strong>and</strong> Benefits<br />

<strong>of</strong> SEL – UNESCO MGIEP conducts an online<br />

presentation <strong>and</strong> discussion on SEL <strong>and</strong><br />

Global Citizenship<br />

July 2018<br />

A key focus area for UNESCO MGIEP is content<br />

development in socio-emotional learning, as part <strong>of</strong> which<br />

<strong>the</strong> Institute delivered an online pre-recorded presentation<br />

on SEL as a key essence <strong>of</strong> achieving global citizenship.<br />

As part <strong>of</strong> <strong>the</strong> presentation, Dr. N<strong>and</strong>ini Chatterjee Singh,<br />

Programme Specialist –Science <strong>of</strong> Learning, posed some<br />

reflective questions to <strong>the</strong> educator community, including:<br />

1) Discuss one compassionate act you did today 2) Can<br />

you think <strong>of</strong> some ways to encourage empathy 3) Discuss<br />

how socio-emotional learning can benefit students’ overall<br />

learning. To view <strong>the</strong> presentation <strong>and</strong> read responses to <strong>the</strong><br />

reflective questions, visit: https://bit.ly/2OvBpkc .<br />

UNESCO MGIEP’s Games for Learning Programme<br />

in collaboration with Fields <strong>of</strong> View conducted <strong>the</strong> first<br />

‘Training <strong>the</strong> Trainers’ Workshop on <strong>the</strong> game Cantor’s<br />

World. Cantor’s World, one <strong>of</strong> <strong>the</strong> many learning products<br />

developed by UNESCO MGIEP, is a simulation based<br />

game developed to underst<strong>and</strong> <strong>the</strong> Inclusive Wealth Index,<br />

<strong>and</strong> to underst<strong>and</strong> <strong>the</strong> uncertainties involved in governing a<br />

complex socio-economic system <strong>of</strong> a country. <strong>The</strong> game is<br />

targeted towards master’s level students <strong>of</strong> economics <strong>and</strong><br />

sustainability studies <strong>and</strong> mid-career policy makers. <strong>The</strong><br />

workshop was conducted as a first step <strong>of</strong> <strong>the</strong> pilot study<br />

to demonstrate <strong>the</strong> game to invited lecturers from different<br />

universities, introduce <strong>and</strong> train <strong>the</strong>m on <strong>the</strong> game, in<br />

order to facilitate its deployment as a pedagogical tool in<br />

classrooms in <strong>the</strong>se universities. <strong>The</strong> lecturers will go back<br />

<strong>and</strong> use this game as a learning resource in <strong>the</strong>ir classrooms<br />

<strong>and</strong> suggest improvements to <strong>the</strong> curriculum modules <strong>and</strong><br />

<strong>the</strong> game itself.<br />

ISSUE • 9<br />

5 5



Dr. N<strong>and</strong>ini Chatterjee Singh accorded <strong>The</strong><br />

NASI – Reliance Industries Platinum Jubilee<br />

Award 2018<br />

Ecology, Economy <strong>and</strong> Society: Essays in<br />

Honour <strong>of</strong> Kanchan Chopra by Vikram Dayal<br />

(Editor), Anantha Duraiappah (Editor),<br />

N<strong>and</strong>an Nawn (Editor)<br />

August 2018, India<br />

UNESCO MGIEP’s Director, Dr. Anantha Duraiappah<br />

recently co-edited a book titled ‘Ecology, Economy <strong>and</strong><br />

Society: Essays in Honour <strong>of</strong> Kanchan Gupta’. Published<br />

by Springer, <strong>the</strong> book includes four vital <strong>the</strong>mes related to<br />

<strong>the</strong> environment-development discourse -- sustainability <strong>of</strong><br />

development, institutions <strong>and</strong> environmental governance,<br />

environment <strong>and</strong> well-being, <strong>and</strong> ecosystem <strong>and</strong><br />

conservation. <strong>The</strong> book is a valuable resource for students,<br />

teachers, researchers, practitioners <strong>and</strong> policy makers. A<br />

Memor<strong>and</strong>um <strong>of</strong> Underst<strong>and</strong>ing (MoU) has been signed<br />

between Samsung India <strong>and</strong> UNESCO MGIEP to launch<br />

360-degree videos <strong>and</strong> VR educational content for 28<br />

UNESCO Heritage sites in India.<br />

#KindnessMatters: Indian youth show<br />

force <strong>of</strong> Kindness across <strong>the</strong> country on<br />

International Youth Day through Instameets<br />

in 14 cities on <strong>the</strong> Sustainable Development<br />

Goals<br />

August 12, 2018, India<br />

More <strong>of</strong>ten than not, youth are <strong>the</strong> subject <strong>of</strong> skepticism<br />

when it comes to stepping out <strong>of</strong> <strong>the</strong>ir comfort zones <strong>and</strong><br />

taking tangible steps towards <strong>the</strong> Sustainable Development<br />

Goals (SDGs). However, on International Youth Day<br />

(Sunday, 12 August 2018) thous<strong>and</strong>s <strong>of</strong> Indian youth<br />

display <strong>the</strong> power <strong>of</strong> kindness across 14 cities joined<br />

forces to perform voluntary Acts <strong>of</strong> Kindness towards <strong>the</strong><br />

Sustainable Development Goals <strong>and</strong> proved naysayers<br />

wrong. <strong>The</strong> unique pan-India Instameets, organised by<br />

UNESCO MGIEP <strong>and</strong> Instagram India, were aimed at<br />

mobilising youth to take tangible action towards realising<br />

<strong>the</strong> Sustainable Development Goals, primarily in <strong>the</strong><br />

form <strong>of</strong> collaborative acts <strong>of</strong> kindness. While <strong>the</strong>se acts <strong>of</strong><br />

kindness encompassed many issues, <strong>the</strong>re was unanimous<br />

commitment by all youth groups across SDGs <strong>and</strong> cities.<br />

<strong>The</strong>se unique Instameets marked <strong>the</strong> India launch <strong>of</strong><br />

<strong>the</strong> International Youth Campaign on Kindness for <strong>the</strong><br />

Sustainable Development Goals, which aims at collecting<br />

25,000 stories <strong>of</strong> kindness by youth across <strong>the</strong> world by <strong>the</strong><br />

end <strong>of</strong> 2019. It is <strong>the</strong> first in <strong>the</strong> line <strong>of</strong> many international<br />

events celebrating youth action towards realising <strong>the</strong> SDGs.<br />


August 2018, India<br />

Dr. N<strong>and</strong>ini Chatterjee Singh, Programme Specialist –<br />

Science <strong>of</strong> Learning, UNESCO MGIEP was recently<br />

accorded <strong>The</strong> National Academy <strong>of</strong> Sciences, India<br />

- Reliance Industries Platinum Jubilee Award 2018 for<br />

Application Oriented Innovations covering both Physical<br />

<strong>and</strong> Biological Sciences. Dr. Singh is one <strong>of</strong> six recipients<br />

<strong>of</strong> this prestigious national award. At UNESCO MGIEP,<br />

N<strong>and</strong>ini oversees <strong>the</strong> Rethinking Learning Programme,<br />

including Libre, our flagship project <strong>and</strong> Difference<br />

Learning, amongst o<strong>the</strong>rs.<br />

An Ahinsa Dialogue on ‘Creating a Culture <strong>of</strong><br />

Peace’ – a conversation on peace, education<br />

<strong>and</strong> neuroplasticity<br />

October 2, 2018 | Paris, France<br />

In commemoration <strong>of</strong> <strong>the</strong> beginning <strong>of</strong> <strong>the</strong> year-long<br />

celebration <strong>of</strong> Mahatma G<strong>and</strong>hi’s 150 th birth anniversary<br />

<strong>and</strong> <strong>the</strong> International Day <strong>of</strong> Non-Violence 2018, <strong>the</strong><br />

UNESCO Mahatma G<strong>and</strong>hi <strong>of</strong> <strong>Education</strong> for Peace <strong>and</strong><br />

Sustainable Development (MGIEP) <strong>and</strong> <strong>the</strong> Permanent<br />

Delegation <strong>of</strong> India to UNESCO jointly organised <strong>the</strong> third<br />

Ahinsa Dialogue on October 2, 2018. <strong>The</strong> dialogue featured<br />

Sadhguru in discussion with Gregoire Borst, Pr<strong>of</strong>essor <strong>of</strong><br />

Developmental Psychology <strong>and</strong> Cognitive Neuroscience<br />

<strong>of</strong> <strong>Education</strong>, LaPsyDE, on ‘Creating a Culture <strong>of</strong> Peace’.<br />

Moderated by Dr Anantha Duraiappah, <strong>the</strong> discussion<br />

involved Sadhguru <strong>and</strong> Gregoire engaging in a conversation<br />

on ‘Peace, Neuroplasticity <strong>and</strong> <strong>Education</strong>’. Over 500<br />

participants attended <strong>the</strong> dialogue, including various<br />

dignitaries, educators, academics, policy-makers <strong>and</strong> youth.<br />

ISSUE • 9<br />

5 7



#KindnessMatters: Youth across <strong>the</strong><br />

globe come out in large numbers to launch<br />

UNESCO MGIEP’s International Youth<br />

campaign on Kindness for <strong>the</strong> SDGs<br />

October 2, 2018 | New Delhi, INDIA<br />

<strong>The</strong> global launch <strong>of</strong> UNESCO MGIEP’s campaign<br />

witnessed mass participation from youth in four countries---<br />

South Africa, India, Pakistan <strong>and</strong> Mexico where UNESCO<br />

MGIEP’s International Youth Campaign on Kindness for<br />

<strong>the</strong> Sustainable Development Goals was simultaneously<br />

launched via a live link connecting <strong>the</strong> four nations.<br />

In true spirit <strong>of</strong> <strong>the</strong> campaign (which aims to mobilise youth<br />

acts <strong>of</strong> kindness towards <strong>the</strong> Sustainable Development<br />

Goals), <strong>the</strong> campaign launch witnessed massive acts <strong>of</strong><br />

kindness committed by youth for one or more <strong>of</strong> <strong>the</strong><br />

Sustainable Development Goals.<br />

#TAG e Kindness: Youth <strong>and</strong> policymakers<br />

dissect <strong>the</strong> role <strong>of</strong> “Kindness” in a time <strong>of</strong><br />

Societal <strong>and</strong> Planetary change<br />

#KindnessMatters: Pakistan youth say ‘yes’<br />

to Kindness; launch MGIEP’s International<br />

Youth Campaign on Kindness at YESPeace<br />

Festival<br />

October 2, 2018 | Islamabad, Pakistan<br />

#KindnessMatters: In a befitting celebration,<br />

South African youth join forces to launch<br />

UNESCO MGIEP’s Campaign on Kindness on<br />

<strong>the</strong> ocassion <strong>of</strong> International Day <strong>of</strong> Non-<br />

Violence<br />

October 2, Cape Town<br />

Africa Unite, in partnership with Activate! Change<br />

Drivers <strong>and</strong> <strong>the</strong> UNESCO MGIEP <strong>and</strong> Sustainable<br />

Development (UNESCO – MGIEP), Western Cape<br />

Provincial Government <strong>and</strong> City <strong>of</strong> Cape Town launched<br />

<strong>the</strong> International Youth Campaign on Kindness for <strong>the</strong><br />

Sustainable Development Goals on October 2, 2018, at <strong>the</strong><br />

Cape Town City Hall.<br />


October 2, 2018 | New Delhi, INDIA<br />

On <strong>the</strong> occasion <strong>of</strong> <strong>the</strong> United Nations International<br />

Day <strong>of</strong> Non-violence, <strong>and</strong> in line with its vision <strong>of</strong><br />

mainstreaming <strong>the</strong> collective voice <strong>of</strong> youth in <strong>the</strong> highest<br />

levels <strong>of</strong> policymaking processes in education for peace,<br />

sustainability <strong>and</strong> global citizenship, UNESCO Mahatma<br />

G<strong>and</strong>hi Institute <strong>of</strong> <strong>Education</strong> for Peace <strong>and</strong> Sustainable<br />

Development (MGIEP) hosted <strong>the</strong> latest edition <strong>of</strong> its<br />

flagship TAG e dialogue series as part <strong>of</strong> <strong>the</strong> Global Launch<br />

<strong>of</strong> <strong>the</strong> International Youth Campaign on Kindness for <strong>the</strong><br />

SDGs in 4 cities: New Delhi, India; Islamabad, Pakistan;<br />

Cape Town, South Africa <strong>and</strong> TEC de Monterrey, Mexico.<br />

Over 130 young members <strong>of</strong> YESPeace Pakistan came<br />

toge<strong>the</strong>r in Islamabad on October 2 to launch <strong>the</strong><br />

International Youth Campaign on Kindness for <strong>the</strong><br />

Sustainable Development Goals (SDGs). <strong>The</strong> launch took<br />

place at <strong>the</strong> Hill View Hotel wherein <strong>the</strong> youth shared<br />

<strong>the</strong>ir kindness stories with each o<strong>the</strong>r <strong>and</strong> video connected<br />

with <strong>the</strong> launch events in New Delhi, Cape Town <strong>and</strong><br />

Monterrey (Mexico). <strong>The</strong> event was organised <strong>and</strong> managed<br />

by YESPeace Pakistan, a collective <strong>of</strong> youth organisations<br />

in Pakistan that included HIVE <strong>and</strong> College <strong>of</strong> Youth<br />

Activism <strong>and</strong> Development (CYAAD). <strong>The</strong> programme was<br />

supported actively by volunteers <strong>and</strong> facilitators from <strong>the</strong><br />

International Federation <strong>of</strong> Red Crescent (IFRC) Pakistan.<br />

ISSUE • 9<br />

5 9

Learning with A.I.<br />

COMIC<br />

CORNER<br />

Seesaws are<br />

so much fun!<br />

I wonder how<br />

<strong>the</strong>y work..<br />

Oh look!<br />

<strong>The</strong>re’s our teacher,<br />

we can ask her.<br />

Miss, when we get back<br />

to class, can you teach us<br />

how a seesaw works?<br />

Why later?<br />

We can learn that<br />

right now using<br />

Artifical <strong>Intelligence</strong>!<br />

See kids, A.I. can<br />

make learning fun<br />

<strong>and</strong> possible anywhere!

<strong>The</strong> <strong>Blue</strong> DOT is UNESCO MGIEP’s bi-annual publication, featuring articles showcasing our activities<br />

<strong>and</strong> areas <strong>of</strong> interest. <strong>The</strong> magazine’s overarching <strong>the</strong>me is <strong>the</strong> relationship between education, peace,<br />

sustainable development <strong>and</strong> global citizenship. To view <strong>the</strong> e-publication, visit –<br />


Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!