Views
2 weeks ago

Diplomatic World_nummer 56.

EU/CHINA/USA – THE

EU/CHINA/USA – THE ARTIFICIAL INTELLIGENCE (R)EVOLUTION – BIG DATA, BIG BUSINESS, BIG CHALLENGES Some reflections on and a summary of a one-day workshop (19 December 2017 - http://www.ucip.be/ Events/), organized by the UCIP 1 project (funded by INNOVIRIS), BACES 2 , BDA 3 and the Confucius Institute at VUB. 104 “Artificial Intelligence” (AI) used to be the privileged playground of universities and research institutes. They were steering the progress in the field. Nowadays this is slightly different. AI has also become a field of competition between large companies and power blocks. Multinational companies have become a major source of progress and innovation, with immense societal and economic impact. AI is also of considerable interest to governments, policy makers and the public at large. Regulations are being designed, sometimes driven by fear, ethical and security considerations but also with the objective of improving the quality of life, both at the societal and the individual level. In the workshop, the recent AI (r)evolution was analyzed and discussed from the point of view of research and development as well as its economic impact. The emphasis was on the China/Europe/USA perspective. Currently, China and the USA are battling to become the world’s first AI superpowers. The USA is still the leading nation, while Europe seems to be absent or at least less visible, in this race. Companies like Facebook, Google, Amazon and Microsoft are being challenged by the likes of Baidu, Alibaba and Tencent (B.A.T.). Analyzing the reasons why, in the future, the use of AI in China could surpass the USA and definitely Europe was a central question in the debate. As AI will impact our lives profoundly in the coming 10 years, are we comfortable buying Chinese products powered with Chinese AI? Would we buy a self-driving car or intelligent fridge made by a Chinese company? Are we OK with having eye surgery or accepting a cancer diagnosis from a Chinese AI robot? How can we best cope with the realities of China’s AI trends, ambitions and globalization plans? How can we assess the benefits and risks of Chinese Intelligence getting into our lives undetected? What is Europe doing to secure a leading role in the AI field and capitalize on university research results? What role can EU- China relations play here? How can we overcome barriers in language and data regulation when collaborating with China? Is there an inherent danger over collaboration with China? How can we address Chinese market opportunities and risks? What is the role of the three big tech companies, B.A.T.? Given the huge size of the Chinese market, how should we adapt our scale of thinking? These are but a few consumer and policy questions that enlightened the debate. Setting the tone for the rest of the workshop, Tias Guns highlighted some aspects of the broader AI context by reminding us of fundamental questions about the nature and the societal impact of the current AI revolution (or is it a bubble 4 ): Is AI a new type of human-created intelligence or is it a research field in computer science that develops smart algorithms? While the first option is still a dream, the second one is reality. Prediction and reasoning have penetrated the world of computer science: decision support systems propose action and autonomous systems act. “Big data” has become a container concept for diverse disciplines in data science and the playground for an evolution from knowing (collecting heterogeneous data to analyze the present — information extraction), prediction (learning from the past to predict the future — machine learning) and reasoning (planning and acting according to the now and predicted future — artificial intelligence). Can AI outsmart people? Yes, a person is limited in space and time while AI is limited in neither because it can rely on parallel computing and the cloud. Is AI. Dangerous? Potentially yes, AI systems are like savants who are unusually good at doing one thing, but with no context, no morality, no knowledge of side effects. Tias Guns concluded his talk with the statement: “AI can make

wishes come true, after investments in infrastructure and data. Let’s hope these wishes are leading to human-oriented, sustainable solutions.” Pascal Coppens gave an inspiring talk on “Artificial Intelligence: China’s new normal”. China’s ambition is to become the leader in AI by 2030. Noteworthy in President Xi Jinping’s speech at the 19th CPC National Congress in 2017 was the following statement: “We will work faster to build China into a manufacturer of quality and develop advanced manufacturing, promote further integration of the internet, big data, and artificial intelligence with the real economy”. In terms of numbers of companies involved, investment and attracting talent, China seems to be well on its way. The Chinese government plans to win the AI race against the USA. Whereas the USA is mainly focused on horizontal generic platforms that provide long term benefits, China is focused primarily on vertical platforms that bring rapid financial returns, e.g. in specific areas such as speech and face recognition, computer vision and intelligent robotic systems. AI technology platforms are being built by the Chinese internet giants, as in the USA. Looking at the key performance indicators — AI companies, AI investments and AI talent — the USA is still ahead with a factor ranging from 1.5 to 3 compared to China, followed by the UK. The Chinese government sees AI as an effective way to solve the huge challenges facing the country, as well as an accelerator for its economic growth. The Chinese consumer is an early adopter of AI-driven products because of China’s particular cultural context, and Chinese attitudes towards privacy, confidentiality, trust, man-machine-relationships and the urge for more convenience. European consumers are more reluctant and hesitant. China will become an AI economic power because it has the four critical ingredients like the USA: availability of data, infrastructure (e.g. supercomputing combined with the cloud), talent and, in the near future, advanced chip designs and implementation facilities. The real disruption in AI will come from the hundreds of well-funded new generations of global Chinese start-ups, especially those creating the latest consumer devices built in Shenzhen. Note of the author: It is unclear if universities will be able to play a broader role than just talent creation for these start-ups. Most of them are not yet well prepared for systematic spin-off creation through wellestablished technology transfer processes. Therefore, also in VUB’s UCIP project, new types of contacts are currently being established in China, directly with match-making companies (IP brokers) and with companies located within the partner university’s ecosystem, bypassing the university’s technology transfer units. Pascal Coppens ended his talk with the unanswered question: But where is Europe? Note of the author: There 105