CS Jul-Aug 2025
Transform your PDFs into Flipbooks and boost your revenue!
Leverage SEO-optimized Flipbooks, powerful backlinks, and multimedia content to professionally showcase your products and significantly increase your reach.
Computing
Security
Secure systems, secure data, secure people, secure business
FIT FOR PURPOSE
New guidance on
storage media
disposal raises
doubts for future
NEWS
OPINION
INDUSTRY
COMMENT
CASE STUDIES
PRODUCT REVIEWS
WARNING SIGNS
Ransomware attacks
offer hope and horror
AI - FRIEND OR FOE?
Is artificial intelligence
to be trusted or is it
just masquerading
as an ally?
BLAST-OFF FOR QUANTUM
Battle to outpace the hackers
now moves into outer space
Computing Security July/August 2025
What if there was a way to
Adapt to all Email
Security threats...
Libraesva integrates cloud email and a secure email gateway with our unique
adaptive trust engine to provide award winning protection.
Layered security defends your business against spam, malware, phishing, email
fraud, spoofing, zero-day threats, account takeover, social engineering, business
email compromise, inadvertent disclosure of sensitive information and more.
Test your security for FREE with our Email Security Tester
emailsecuritytester.com
libraesva.com
comment
NHS CASH INJECTION CAN'T QUELL ATTACK FEARS
EDITOR: Brian Wall
(brian.wall@btc.co.uk)
LAYOUT/DESIGN: Ian Collis
(ian.collis@btc.co.uk)
SALES:
Edward O’Connor
(edward.oconnor@btc.co.uk)
+ 44 (0)1689 616 000
David Bonner
(dave.bonner@btc.co.uk)
+ 44 (0)1689 616 000
Stuart Leigh
(stuart.leigh@btc.co.uk)
+ 44 (0)1689 616 000
It's encouraging to see that the NHS received a record cash investment in the UK Spending
Review, with an additional 10% technology budget increase.
But the question being asked in many quarters is: does this investment include cyber security
defence for one of our most vulnerable critical infrastructures? It is widely recognised that the
healthcare sector, dealing with sensitive personal information and running largely on legacy
systems, is highly attractive and vulnerable to cyber-attackers.
Reports reveal that only 36% of NHS staff believe cyber security measures are sufficient. Data
from SonicWall reflects this, showing healthcare remains a prime target, with hospitals and
healthcare systems taking 60-150 days to patch vulnerabilities, while hackers exploit them in just
two days, creating a critical security gap.
Executive VP EMEA Spencer Starkey at SonicWall, says that, despite Rachel Reeves' NHS budget
increase, the healthcare sector remains a prime target for ransomware attacks, as healthcare
organisations often have critical data that they cannot afford to lose. "If they haven't already,
healthcare organisations need to have a plan in place to respond to ransomware attacks and to
minimise the impact of these attacks. The healthcare sector continues to be a prime target for
malicious actors, as evidenced by the recent attacks on the NHS," he points out.
"Not only do these attacks risk the potential for exposed patient data, but any significant IT issue
that halts patient care poses an immediate threat to life," adds Starkey. "The ramifications of an
attack on the healthcare sector can be disastrous, and it's important to place the utmost amount
of time, money and efforts on securing them."
Brian Wall
Editor
Computing Security
brian.wall@btc.co.uk
PUBLISHER: John Jageurs
(john.jageurs@btc.co.uk)
Published by Barrow & Thompkins
Connexions Ltd. (BTC)
Suite 2, 157 Station Road East
Oxsted. RH8 0QE
Tel: +44 (0)1689 616 000
Fax: +44 (0)1689 82 66 22
SUBSCRIPTIONS:
UK: £35/year, £60/two years,
£80/three years;
Europe: £48/year, £85/two years,
£127/three years
R.O.W:£62/year, £115/two years,
£168/three years
Single copies can be bought for
£8.50 (includes postage & packaging).
Published 6 times a year.
© 2025 Barrow & Thompkins
Connexions Ltd. All rights reserved.
No part of the magazine may be
reproduced without prior consent,
in writing, from the publisher.
www.computingsecurity.co.uk July/August 2025 computing security
@CSMagAndAwards
3
Secure systems, secure data, secure people, secure business
Computing Security July/August 2025
inside this issue
CONTENTS
Computing
Security
NEWS
OPINION
INDUSTRY
COMMENT
CASE STUDIES
PRODUCT REVIEWS
FIT FOR PURPOSE
WARNING SIGNS
Ransomware attacks
New guidance on
offer hope and horror
storage media
disposal raises
doubts for future
AI - FRIEND OR FOE?
Is artificial intelligence
to be trusted or is it
just masquerading
as an ally?
BLAST-OFF FOR QUANTUM
COMMENT 3
NHS: will new cash-in stop it crashing?
Battle to outpace the hackers
now moves into outer space
NEWS 6
Imposter in high-level AI hoax
Partnering up to tackle risk
JumpCloud Acquires VaultOne
Data leak exposes 5.7 million files
ARTICLES
QUANTUM QUAKES 18
The incredible processing power and
speed of quantum computers significantly
threaten traditional encryption methods.
How do we guard against this?
GOOD, BAD AND WAY FORWARD 10
Ransomware attacks are said to have
declined by 23% between April-June this
year, compared to the previous quarter.
Meanwhile, industrials are still the most
targeted sector, with attacks up 46%
from Q4 2024 to Q1 2025, according to
a new threat report.
HEALTH ISSUES 20
The NHS has received a record cash
investment in the latest UK Spending
SQUARING THE CIRCLE 14
Review. Will the extra injection prove
Fresh top-level guidance on the secure
enough to revive it?
disposal of storage media may raise more
questions than it answers, says one industry
RETAIL AND RETALIATION 21
expert. Missing are familiar terms such as
The recent wave of attacks on retail
'Clear' or 'Purge' and there is no reference to
companies has shown no organisation,
the IEEE 2883:2022 sanitisation specification,
whatever their status might be, is safe
a key industry standard.
from those who target their operations
DATA THEFT ONSLAUGHT 25
Four people were arrested in raids over the
wave of cyber-attacks that crippled M&S,
SEVEN-YEAR ITCH 22
the Co-op and Harrods. But this won't
stop more such assaults taking place
It's been more than seven years now since
the General Data Protection Regulation
CALL FOR STRONGER CYBER LAWS 30
(GDPR) came into force. With the rapid
MPs have shown widespread support for
acceleration of AI technologies and the
stronger cyber laws, along with calls for
increasing use of big data to train models
greater government collaboration and a
and automate decisions, can the GDPR
more ambitious, future-proofed approach
keep pace and remain fit for purpose?
16 BILLION REASONS FOR ACTION 32
Data breaches that resulted in a massive
16 billion passwords being stolen could
AI - FRIEND OR FOE? 26
create a snowball effect of cyber-attacks
AI systems are all about the quality of their
in the days ahead
training data, one observer tells Computing
QUANTUM BLAST-OFF! 34
Security. Models fed narrow, outdated or
Researchers in Europe and Canada are
even fundamentally biased datasets may
developing technology that will allow them
overlook new threats or reproduce the bias
to fire quantum-coded messages across
at speed and scale. Adversaries are quick to
continents via satellites in space.
exploit these cracks, he warns.
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
4
Layers aren’t just for cakes; they’re
essential in cybersecurity’s secret
recipe for protection!
Bake it happen with VIPRE Security Group. Secure your
bytes before you take a bite with Email Security, Endpoint
Security and User Protection
www.vipre.com
news
Jason Soroko, Sectigo.
MARCO RUBIO IMPOSTOR USES AI TO CALL
HIGH-LEVEL GOVERNMENT OFFICIALS
An imposter has used AI to impersonate US secretary
of state Marco Rubio in calls to high-ranking
government officials.
The threat actor employed AI-powered software to
mimic Rubio's voice and writing style, with the intention
of manipulating foreign ministers of unnamed countries.
The incident highlights the way that AI is enabling
constantly evolving novel methods of carrying out cyberattacks.
Spencer Starkey,
SonicWall.
"The most effective method to safeguard our systems in
this scenario would be to leverage AI to compete against
its own potential threats in real-time," says Spencer Starkey, executive vice president
EMEA, SonicWall. "This perspective introduces a novel concept: a cybersecurity landscape
where AI engages in a continuous battle against cyber threats. Unlike traditional warfare,
this battlefield knows no seasons or holidays; it is a relentless, 24/7 endeavour to protect
our digital assets."
MAJOR MOVE IN QUANTUM
SUPREMACY RACE
Sir Jeremy Fleming, the former head
of GCHQ, has joined the board of
Oxford quantum computing start-up
Oxford Quantum Circuits.
PARTNERING UP TO TACKLE RISK
Sophos has formed a new partnership with Capsule,
a specialist insurance broker, that facilitates access
to cyber insurance coverage for organisations
deploying Sophos' cybersecurity solutions via a
managed services provider (MSP).
Liam Green, Capsule
This is seen as a major and significant
development as Britain races against
China and the US, with plans to spend
£2.5bn on quantum computing in
pursuit of a competitive edge.
As the UK sets out its bold roadmap
to build a world class quantum
workforce, however, it's essential we
keep quantum security front and centre,
cautions Jason Soroko, a security technology
innovator and senior fellow at
Sectigo. "This report highlights the
transformative potential of quantum
technologies, but does not pay enough
attention to the potential risks that
quantum computing brings to the
cybersecurity space - with only a token
reference made to the possible creation
of 'Quantum cryptography/cyber
security' roles.
"Sophos users enjoy automatic premium reduction, a
streamlined application process, comprehensive
coverage and pre-approved use of Sophos incident
response services, while Sophos MSPs are better able
to support their customers with a trusted cyber
insurance solution," states the company.
Adds Liam Green, co-founder and chief operating officer at Capsule: "Cybersecurity
and cyber insurance can no longer operate in silos - they must work together to
create measurable risk reduction for businesses."
GROWING GAINS
Advania UK, one of Microsoft's leading partners in Northern Europe, has
announced powerful growth and expansion. "The company achieved a record
turnover of more than £450 million and quadrupled its client base, while
significantly expanding its workforce," it states.
Advania's combined turnover for the new UK group was £452 million last year,
which was 3.3x higher than reported turnover for Advania UK the previous year.
6
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
Building cyber security
awareness together.
Leading the way in personalised
cyber security awareness.
Keep your staff engaged, cyber-secure, and compliant with our award-winning,
personalised cyber security training.
Designed with real people and teams in mind, our expertly crafted content transforms
cyber security into an informative and captivating experience. By making learning
fun and impactful, we maximise engagement and enhance staff security behaviour,
ensuring constant vigilance against cyber threats.
Our staff fully engaged with our
security awareness program, with
completion rates over 85%
Best cyber security awareness
platform available
news
Greg Keller,
JumpCloud.
KEY APPOINTMENT AT HACKERONE
HackerOne, a global leader in offensive security solutions,
has appointed Nidhi Aggarwal as chief product officer and
member of the executive leadership team.
"Aggarwal will lead the execution of HackerOne's platform
vision and product strategy, unifying the company's product
portfolio around a more integrated, AI-powered experience that
seamlessly scales human security expertise through AI agents to
not just find, but remediate vulnerabilities," says the company.
Nidhi Aggarwai,
HackerOne.
A seasoned technology entrepreneur and product leader, she
brings more than 15 years' experience in driving growth and
innovation at companies ranging from early-stage startups to
global enterprises.
JUMPCLOUD ACQUIRES VAULTONE
JumpCloud sees its recent acquisition of
VaultOne as a step on the path towards
igniting a new era of privileged access
management.
"We are continuously evaluating our
customers' needs to extend JumpCloud into
more areas where we can provide secure,
frictionless access to resources, regardless
of where the team or resources are
located," says Greg Keller, chief technology
officer and co-founder, JumpCloud.
"The acquisition of VaultOne brings a
deeply experienced team and established
PAM technology into the JumpCloud
family. This allows us to immediately offer
a dedicated PAM solution with the granular
control necessary to navigate today's
complex security landscape, while
providing more critical access control
capabilities our customers can consolidate
and depend upon from JumpCloud's
platform."
Stated benefits from the move include:
privileged access for all an organisation's
critical assets; secure browsing, built-in;
removal of the need for VPN; and the
radical reduction of cyber threats.
DATA LEAK EXPOSES 5.7 MILLION FILES
Cybernews researchers recently uncovered a massive data
leak, which was traced back to HireClick, a recruitment
platform for small to mid-sized businesses. The platform
helps businesses manage job listings, candidate applications
and the hiring process.
The company left over 5.7 million files wide open for anyone
on the internet, thanks to a misconfiguration of Amazon
AWS S3 storage bucket. The leaked files exposed sensitive
and private information of job seekers, mainly resumés.
"In the wrong hands, the leaked data could power everything
from identity theft and impersonation to phishing, vishing,
and smishing campaigns, where attackers pose as hiring managers to exploit desperate job
seekers," warns Cybernews journalist Paulina Okunytè.
KITEWORKS EXPANDS MARKET REACH
Amit Toren, Kitgeworks.
Kiteworks has acquired Zivver, a secure email platform
headquartered in Amsterdam, the Netherlands.
"Organisations require comprehensive solutions that
protect private data shared and sent across all
communication channels while maintaining seamless
user experiences," says Amit Toren, chief business officer
at Kiteworks. "By bringing Zivver's innovative secure email
technology into our Private Data Network, we're
providing the combined customer base with enhanced
capabilities to secure their most sensitive
communications, while meeting stringent compliance
requirements.
Paulina Okunytè
8
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
ransomware
WHAT DIRECTION NOW AS
RANSOMWARE HITS HARD?
RANSOMWARE REMAINS A FORMIDABLE ADVERSARY, ESPECIALLY WHEN
ITS TARGETS ARE THE BACKBONE OF NATIONAL INFRASTRUCTURE
First, the relatively good news.
Ransomware attacks are said to have
declined by 23% between April and
June this year, compared to the previous
quarter, although they are up 43% on this
time last year, with the dip only partially
explained by normal seasonal variations.
"In Q2 of 2025, 1,591 new victims of
ransomware attacks were posted publicly
on data leak sites, at an average of 17.5
per day, compared to 22.9 per day in Q1
of 2025 and 12.2 per day in Q2 of 2024,"
reports Steve Alder, editor-in-chief of The
HIPAA Journal.
Now, the bad news: industrials as still
being the most targeted sector, showcasing
the value that critical national infrastructure
holds for ransomware groups. Attacks
jumped by 46% from Q4 2024 to Q1 2025,
according to Honeywell's 2025 Cybersecurity
Threat Report. The research also found that
both malware and ransomware increased
significantly in this period and included a
3,000% spike in the use of one trojan
designed to steal credentials from industrial
operators. Where do these differing takes
leave organisations in their quest to stay
'ransom free'?
A COMPLEX REALITY
Daniel Shepherd, CEO at CSIS Security
Group, says recent research offers a
nuanced view of ransomware. "While several
reports, including findings from the NCC
Group, suggest a decline in overall attack
volumes, this improvement should not lull
us into a false sense of security. Underneath
these encouraging figures lies a more
complex reality. Ransomware remains
a formidable adversary, especially when
its targets are the backbone of national
infrastructure."
The most recent CSIS Threat Matrix report
reinforces this cautionary note, he says,
emphasising the urgent need for resilient,
targeted countermeasures in the face of
a threat that continuously adapts and
intensifies. "Even if the frequency of reported
incidents has momentarily dipped, adversaries
are still honing their techniques to
exploit weaknesses. Now, ransomware
actors are not simply relying on traditional
encryption to lock down data, they are
employing double extortion tactics that
threaten both operational integrity and
corporate reputation.
"This sophisticated hybrid of technical
prowess and psychological pressure means
that any lull in activity can be misleading.
Organisations across all sectors must
remain vigilant, because complacency is
the attacker's greatest ally. The need for a
proactive, dynamic cybersecurity stance has
never been more critical," warns Shepherd.
Drawing on insights from the latest CSIS
Threat Matrix report, a multi-layered strategy
is essential for thwarting ransomware
attacks. Here are the key pillars to consider,
he suggests:
Comprehensive risk assessments and
continuous monitoring
"Organisations need to carry out frequent
and detailed vulnerability audits to uncover
system-specific weaknesses. Leveraging
advanced endpoint detection and response
(EDR) tools can help detect suspicious file
modifications and unusual encryption
surges; both of which are signs that a
ransomware attack may be underway."
Robust incident response and threat
intelligence
"The value of preparedness cannot be
underestimated. Develop a dedicated
ransomware response plan that outlines
rapid isolation measures, sets clear communication
roles and includes established
protocols for involving law enforcement
when necessary. Additionally, forging
collaborative intelligence-sharing channels
with industry peers provides invaluable realtime
insight into emerging threats and
evolving tactics."
Employee empowerment through training
"The human element remains both a
vulnerability and a defence," adds Shepherd.
"Tailored scenarios and simulated attack
exercises cultivate a culture of cyber hygiene,
equipping employees to recognise phishing
attempts and other common precursors to
ransomware."
Layered technical defences
A defence-in-depth approach is important,
he says. "Implement robust firewalls, segmented
networks, multi-factor authentication
and behaviour-based anomaly
detection systems. Proactive measures, such
as advanced email filtering, sandboxing
and a rigorous patch management policy,
will significantly reduce exposure to
potential exploits."
10
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
ransomware
Resilient business continuity planning
"Establishing an effective data backup
strategy is a safeguard against crippling
downtime. Ensure backups are immutable,
regularly tested and stored securely. Developing
and rehearsing rapid recovery
protocols allows organisations to bounce
back swiftly, minimising both operational
disruption and financial losses."
BREAKING THE BANK
For a period of time in his military
intelligence career, Rob Dartnall, CEO,
SecAlliance, was focused on disrupting the
finances of terrorists. "What struck me then
was just how much that money influenced
the battlefield. A lesson that continues to
resonate when discussing ransomware
payments. At times, removing the funding
had a bigger impact than kicking down
doors. At times, we would come across
fighters who were not there for ideology,
but simply to put food on the table. No
money, fewer fighters. No money, fewer
weapons. No money, reduced capability.
Starve a group of funding and their longterm
development is significantly restricted.
"This mindset followed me as I transitioned
into cyber security over a decade ago. We're
facing increasingly capable ransomware
actors, with operations that reflect the
structure and scale of traditional adversaries.
The solution remains the same; we must
disrupt the flow of money. That's why I
support recent calls from cyber leaders like
Ciaran Martin [former CEO of the National
Cyber Security Centre] to ban ransom
payments outright. The rationale is clear.
Criminal groups continue to extract
significant sums from public and private
institutions. These funds sustain further
attacks, attract new actors, drive illicit
economies, empower corrupt regimes
and support geopolitical instability."
Dr Richard Horne, current CEO of the
National Cyber Security Centre, part of
GCHQ, has publicly stated his own
opposition to making such payments. "If
[companies] pay because they hope the
ransom attackers won't publish information,
well, they need to know that all they've got
is a criminal's word for it," said Horne. "And
if they pay to recover their systems, well,
they should have recovery plans in place so
they can recover their systems anyway."
Meanwhile, there are signs that coordinated
deterrence has an effect, adds
Dartnall. "Countries and jurisdictions where
ransom payments are restricted, particularly
in cases involving sanctions or terrorism, are
becoming less appealing targets. And when
payments are blocked entirely, threat actors
tend to shift focus." He refers to the recent
international operation against the LockBit
group, led by the UK's National Crime
Agency and its allies, which was not only
a technical achievement, but also
psychological in nature.
TACTICS TURNED ON THEIR HEADS
"Their defeat was carried out using their
own tactics. That's effective strategy:
undermining morale, disrupting
communication and cutting off financial
reward. What has been more impressive is
the sustained law enforcement operations
not only against other nefarious actors,
but also against the markets, infrastructures
and systems they rely on to illicit their gains.
When terrorist groups lose funding, their
operations contract. They make more mistakes
and become easier to detect. The
same pattern applies in cyber."
Apart from banning payment, he says
tackling the infrastructure that facilitates
cybercrime is vital. "Many financial
institutions are already obliged to flag
suspicious transactions, so why not apply
similar standards to cryptocurrency
exchanges - globally. Sanction the owners
of illicit exchanges and markets? These
platforms have too often provided cover
for laundering criminal gains. Regulation
here is overdue."
Daniel Shepherd, CSIS Security Group:
ransomware remains a formidable adversary.
Rob Dartnall, SecAlliance: starve a group of
funding and their long-term development is
significantly restricted.
www.computingsecurity.co.uk @CSMagAndAwards July/August 2025 computing security
11
ransomware
Saugat Sindhu, Wipro: prevention
capabilities should be the cornerstone
of any cyber resilience strategy
Ian Robinson, Titania: boosted by AI,
ransomware is infiltrating all the way
into the network.
However, a blanket ban could also have
unintended consequences, such as
redirecting criminals more towards private
individuals, he recognises. "The outcome
could be damaging, if skilled attackers pivot
to widespread financial fraud against the
public. That's why any response needs to
be coordinated.
Government departments, ISPs, telecom
providers, domain registrars and email
platforms all have a role to play. For the last
five years, the changes in the geopolitical
landscape have completely fractured
relationships between nations that had
shown progress in working together
to eradicate cyber-criminal activity. Whilst
there will be safe harbours for actors, we
can significantly reduce their freedom of
movement, and access to markets and
infrastructures." Prohibiting ransomware
payments isn't a catch-all fix. "It must be
backed by long-term investment, increased
resourcing for law enforcement, improved
cyber defence responsibilities for service
providers and, most importantly, a unified
national strategy. As new technologies allow
for more threat actors to more easily enter
criminal markets, we must similarly restrict
access to resources."
TOP RISK
Saugat Sindhu, global head strategy and
risk at Wipro, points to the company's latest
State of cybersecurity report, which found
that 57% of respondents viewed ransomware
attacks as their top risk. "When it
comes to cyber threats like ransomware,
prevention capabilities should be the
cornerstone of any cyber resilience strategy,
followed by recovery capabilities," says
Sindhu. "The smartest organisations also
learn from the past. They treat threat
reports and breach post-mortems not as
headlines, but as playbooks. Each one
offers clues about what today's attackers
are targeting and how to stay a step ahead.
Every incident adds to existing recovery
SOPs [Standard Operating Procedures]."
Ultimately, the goal is to detect and defuse
threats before they escalate and, if they end
up escalating, quickly recover from them.
"That means identifying vulnerabilities early,
remediating quickly and constantly adapting
defences to meet evolving risks. Cyber resilience
isn't a box to check, it's an ongoing
discipline that is iteratively evolving through
lessons learned."
AI AND AUTOMATION
While attackers traditionally relied on manual
methods, they are now using AI and
automation to accelerate and scale their
efforts, states Ian Robinson, chief product
officer, Titania. He quotes Gartner's
prediction that, by 2027, AI will reduce the
time it takes to exploit account exposures
by 50%. "Boosted by AI, ransomware is
infiltrating all the way into the network.
Instead of immediately infecting a device,
attackers are now deploying more covert
strategies - lying dormant, spreading
undetected and establishing a strong
foothold in the network before executing
the ransomware to maximum effect."
This shift in tactics significantly increases
the potential for business disruption - or
disaster. "People remain the weakest link in
cybersecurity. According to Verizon, human
actions or inactions contributed to 74% of
breaches last year. While organisations must
continue to strengthen their human firewall,
the reality is that it's a matter of when - not
if - they will face an attack. AI-powered
phishing and social engineering are becoming
increasingly convincing, making this
challenge even more pressing."
CRUCIAL CONTROL
Although millions of ransomware variants
exist, they all infiltrate and move through
networks in similar ways, he points out.
This makes network segmentation a crucial
control for mitigating risks, especially for
CNIs [critical national infrastructures].
"If a ransomware attacker breaches the
perimeter, but the administrative network
12
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
ransomware
is segmented from business-critical segments
and operational technology, the
attack will only proliferate as far as it can
get," says Robinson. "This is not the case
for flat net-works that will fail to stop lateral
movement. Lateral movement is an increasing
risk for businesses as attackers use AI
to automate and accelerate credential theft
attacks to open privilege escalation."
Attackers will always look for the most
easily exploitable vulnerabilities to gain
access to systems and/or networks, he adds.
"Network devices are particularly attractive
targets for ransomware, as they provide
attackers with persistent, lateral movement
capabilities and access to sensitive data
flows when compromised. Hardening
these components to enforce network
segmentation is key for CNIs to develop
operational resilience.
"Within CNI, where operational downtime
is not an option, there is often a need to
enable insecure configurations to support
legacy software and operational technology
devices. Out-dated network devices expose
CNIs to the Critical Vulnerabilities and
Exposures [CVEs] repeatedly exploited by
ransomware gangs and APTs [advanced
persistent threats]."
ESSENTIAL SAFEGUARDS
When operating with accepted risks is
unavoidable, foundational network security
and compliance measures are essential to
manage these risks effectively.
This includes minimising the attack surface by
enforcing access control lists and macro
segmentation policies; and proactively monitoring
for device configuration changes to identify
whether a change was planned versus
unplanned, and if it violates segment-ation
policy. By implementing proactive network
security measures, organisations have the realtime
information that an attack has been
mounted and the time to respond and contain
attacks before they become catastrophic."
IMPOSSIBLE CHOICE
AJ Thompson, CCO at Northdoor, says the
M&S attack, which forced a six-week pause
in online operations and an estimated
£300m profit reduction, has demonstrated
the devastating impact that goes so much
further than 'just' data loss. "Organisations
face an impossible choice: adhere to best
practices or try to mitigate the catastrophic
day-to-day business disruption, as well as
the long-term impacts of non-compliance
and associated fines. Our experience
indicates that preparation is the key.
"Companies with robust incident response
and business continuity protocols, immutable
backups and segmented networks give
organisations a strong negotiating leverage,
in the event of a crisis. For those without
such preparation in place, the prospect of
having to pay the ransom demanded by
cybercriminals becomes a very real on. It
might represent the only viable path back
to operational continuity."
There are, however, some key considerations
that will need to be taken into
account, he comments. These include:
The maturity of existing recovery
capabilities
Comprehensive impact assessment
beyond IT systems
Verification of threat actor credibility
through intelligence
Legal and regulatory obligations
regarding data protection.
"The nature of cybercriminals, of course,
means that there is no guarantee that you
will receive all of the stolen data back or in
what form it arrives, but without preparation
it remains the only hope of some to
ensure that their business can get back up
and running. The true solution lies not in
absolute positions, but in proactive resilience
- implementing advanced threat detection,
regular recovery testing and cyber insurance,
with expert response teams to ensure
organisations never face this impossible
choice unprepared."
PROACTIVE STANCE
To mitigate the many threats that ransomware
poses, Lorri Janssen-Anessi, director of
external cyber assessments at BlueVoyant,
argues that "CNI organisations must take a
proactive stance towards cyber security best
practice", with internal teams working with
dedicated MSSPs to:
Regularly back up critical data and
systems, with backups stored offline
to prevent them from being targeted.
Backup and restoration procedures
should be regularly tested to ensure they
are effective in a real-world incident.
Create, maintain and regularly exercise
a cyber incident response plan that
includes specific procedures for ransomware
attacks. This plan should be tested
regularly to identify gaps
Implementation of IAM:
Implement MFA for all services,
especially for remote access, VPNs and
webmail
Restrict user and administrator
privileges to only what is necessary for
their roles
Adopt a zero-trust model, which
assumes no user or device is trusted by
default, requiring strict verification for
every access request.
Keep all operating systems, software and
firmware up to date to protect against
known vulnerabilities that ransomware
actors frequently exploit
Isolate critical systems by segmenting the
network, which can prevent ransomware
from spreading laterally from a
compromised workstation to critical
servers
Educate employees to recognise and
report phishing attempts, suspicious links
and social engineering tactics
Increase detection and response
capabilities to boost overall resilience
against ransomware attacks.
www.computingsecurity.co.uk @CSMagAndAwards July/August 2025 computing security
13
asset disposal
SQUARING THE CIRCLE
FRESH TOP-LEVEL GUIDANCE ON THE SECURE DISPOSAL OF STORAGE MEDIA
MAY RAISE MORE QUESTIONS THAN IT ANSWERS, SAYS ONE INDUSTRY EXPERT
The government's recently launched
'IT Reuse for Good Charter' is widely
seen as a major step forward in
encouraging firms to embed circular
approaches into their daily IT operations -
but it could and should go further still,
says Green Alliance's Emily Carr
"It shouldn't still be the case that, in
the UK today, millions of people have no
access to digital devices to get online,
while perfectly usable business laptops,
tablets and smartphones are routinely
shredded or left gathering dust, adding
to the UK's shocking and ever growing e-
waste problem."
There's one simple solution that tackles
both problems in one go, she believes:
redistributing discarded usable devices
to the people who need them most.
"Businesses are a big part of this solution.
Some are already leading the way,
building this into their IT management
and showing what's possible when
sustainability and social impact are taken
seriously. The recent launch of the 'IT
Reuse for Good Charter' is a major step
forward in helping others to follow their
lead, making device reuse the norm rather
than the exception." Meanwhile, the
NCSC (National Cyber Security Centre)
has published new guidance to help
organisations looking to securely
decommission end-of-life digital assets.
Here are the main points:
Safely retiring data, software and
hardware is a critical endeavour with
"potentially severe repercussions", if
not done right
IT assets allowed to continue beyond
their lifespan may pose a risk to the
organisation, if they are lost, exploited
or accessed by unauthorised individuals
The goal is to understand the potential
impact of the asset's decommissioning
and ensure all associated components
are accounted for.
But how exactly do organisations faced
with such challenges get this right -
and what are the key steps they need
to take to pinpoint the pitfalls awaiting
the unwary?
GAP IN TOP-LEVEL GUIDANCE
"The recent update from the NCSC on
Secure Sanitisation and Disposal of
Storage Media was a welcome
development," says Steve Mellings,
founder and CEO of ADISA Group.
"Previous guidance was spread across
several documents, notably Information
Assurance Standard 5 (IS5), which
provided detailed sanitisation specifications.
This standard operated alongside
the Commercial Product Assurance (CPA)
scheme, which evaluated and approved
sanitisation products, and CAS-S, a certification
mechanism for companies offering
secure disposal services. Together,
these frameworks offered organisations
a high degree of assurance when
disposing of data-bearing assets.
"However, IS5 was last updated in
2014. The CPA scheme ceased approving
sanitisation products in 2023 and CAS-S
has since relied solely on referencing
NPSA destruction standards as an
approved method of sanitisation.
These developments left a noticeable
gap in authoritative guidance."
As Mellings reviewed the new update,
the first thing that stood out for him
was a prominent disclaimer: 'Note: This
guidance will not protect data from being
read by a skilled, well-funded laboratory.'
14
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
asset disposal
This sets a cautious tone, he states. "The
document is also explicitly tied to the UK
protective marking scheme, limiting its
relevance to data classified as OFFICIAL
within HMG. For non-government users,
this raises immediate questions about
applicability and relevance."
The guidance does contain some sound
principles on how organisations might
approach secure sanitisation, he accepts.
"However, what's notably missing is concrete
advice on how to sanitise media.
Familiar terms such as 'Clear' or 'Purge' are
absent and there is no reference to the
IEEE 2883:2022 sanitisation specification,
a key industry standard."
Beyond these omissions, some of the
recommendations are questionable, he
feels. "For example, the suggestion to
'Power off the device for at least 15
minutes' lacks a technical basis in secure
data disposal for most storage media.
Other points - such as reliance on untested
manufacturer resets or acknowledging
that 'data may remain on the device' -
further erode confidence in the document's
technical rigour."
Ultimately, concludes Mellings, this
update feels incomplete and lacking in
real, practical advice. "Rather than closing
the gap left by deprecated standards, it
introduces new uncertainties. For those
seeking clarity and assurance in the secure
disposal of storage media, this guidance
may raise more questions than it answers."
FAILURE OF SCRUTINY
When decommissioning digital assets,
one of the most overlooked risks is the
persistence of sensitive data in unexpected
places, says Richard Hall, AVP at DigiCert.
"While wiping or destroying hard drives
is standard practice, many organisations
fail to apply the same scrutiny to embedded
or secondary storage within devices.
A few years ago, I purchased decommissioned
data centre servers for a home
lab. During set-up, I discovered SD cards
still inserted, containing a full operating
system and boot configuration. It was
a clear reminder that critical data can
remain accessible long after equipment
is retired."
In another case, a colleague had received
a second-hand corporate phone that
had not been properly sanitised. "Despite
a factory reset, the device retained autofill
credentials, saved Wi-Fi networks and
identity tokens, still granting access to
cloud apps. Devices that sync across
platforms often store data in ways that
survive basic resets.
"Network infrastructure is no exception.
Equipment like routers and firewalls can
store credentials, cryptographic keys and
digital certificates in onboard memory.
If left behind, these artefacts can expose
organisations to serious risks - whether
through unauthorised access or persistent
trust relationships."
In today's complex IT environments, adds
Hall, digital certificates are everywhere: from
authentication and encryption to device
identity. "Without oversight, orphaned
certificates on decommissioned systems
can become security and compliance
risks. Many were issued with long validity,
sometimes a decade, during which cryptographic
standards may shift, algorithms
weaken and browser behaviours change.
Without visibility, outdated certificates
can remain active on systems long since
retired."
This highlights the importance of understanding
where cryptographic assets
are deployed, so they can be updated
or retired in line with changing
requirements, he points out. "Two key
developments make this even more
urgent: the shortening of TLS certificate
lifespans to 47 days and the longer-term
need to support post-quantum cryptography
(PQC). An effective decommissioning
strategy must go beyond simply
wiping drives."
Hall says the strategy should include
the following:
Auditing all devices for stored
data and credentials
Revoking certificates tied to
retired systems
Secure erasure or destruction
of all storage
Documenting the process for
compliance and audits.
"When overlooked, these gaps can lead
to data breaches, unauthorised access and
non-compliance. Treated strategically,
decommissioning strengthens digital trust
and operational resilience."
ORDER OF PRIORITY
Decommissioning tends to be regarded
as an end process concerned with the
retirement of IT data, hardware and
software when, in fact, it should be a key
consideration when buying those assets
in the first place, argues Jon Fielding,
managing director, EMEA, at Apricorn.
"The NCSC states that it should be part
of the procurement process but doesn't
explain why. When it comes to physical
storage media, for instance, choosing a
device that is tamper resistant and has
identity-based authentication as demanded
under the FIPS Level 2 and 3 security
standards can significantly reduce risk. If
the company merges or changes hands,
for example, and those devices fall off the
radar, the data is at less risk of being
exploited or accessed by unauthorised
persons."
Rather than regarding decommissioning
as an end process, says Fielding, storage
media should be a trackable asset that is
documented and inventoried from the
get-go. When it needs to be swapped out,
www.computingsecurity.co.uk @CSMagAndAwards July/August 2025 computing security
15
asset disposal
Steve Mellings, ADISA: some of the
recommendations in new guidance on
the secure disposal of storage media are
questionable.
Richard Hall, DigiCert: one of the most
overlooked risks is the persistence of
sensitive data in unexpected places.
either in whole or in part, this needs to be
planned. "A key consideration is not just
where and how the data will be moved,
but also how it will be safeguarded via
backup. Things can and do go wrong
during decommissioning; situations can
change or vulnerabilities emerge, all of
which can necessitate a rollback. So,
making sure there are additional copies
and that a tried and tested recovery plan
is in place, is critical."
To safeguard against these issues, he
cites the '3-2-1 backup rule', which should
be observed. This is where at least three
copies of data are created, with two of
these stored on different media, one of
which should be offsite. "For example, one
copy of the data could be offline on an
encrypted removable hard drive that is
disconnected from the network."
Finally, there's the issue of sanitisation
which should also be a consideration at
the procurement stage particularly as this
can be cost issue. "Encrypted devices, such
as those that use AES 256-bit encryption,
are far more economic to retire as well
as being more secure," he points out.
"This is because it's possible to delete the
encryption keys, thereby rendering the
data unreadable. In contrast, wiping
devices that don't use encryption is a
lengthy process involving overwriting
memory, checking the device metadata,
depowering the device and then checking
the user-accessible memory for any traces
of data. This still follows, even If the
device is being destroyed, as the data
would still need to be removed prior to
destruction."
VECTOR FOR SECURITY THREATS
According to the Eclipse Foundation,
open-source software is rudimentary to
global digital infrastructure and how we
manage the lifecycle of these projects,
particularly their end-of-life (EOL) phase,
carries significant implications. "When an
open-source project is deprecated or
abandoned without a clear EOL strategy,
it can become a vector for serious security
threats. Chief among these is the risk of
exploitation by unauthorised actors, including
package takeover, Git repository
hijacking and even DNS compromise.
Such vulnerabilities have led to real-world
incidents, where attackers republish
malicious versions of libraries that
unsuspecting developers and systems
integrate without scrutiny."
The risk intensifies when maintainers
walk away without formally archiving
repositories, revoking access credentials,
or updating metadata in package
registries, adds the foundation. "In such
a vacuum, attackers can impersonate
former maintainers or exploit forgotten
infrastructure to inject malicious code
into the supply chain. Beyond security,
organisations that originally developed
or were publicly associated with these
projects may suffer reputational damage,
if their abandoned code is compromised,
even if they're no longer directly involved."
At the Eclipse Foundation, all projects
are said to operate within a governance
model that enforces identity verification,
controlled access to repositories, welldocumented
metadata and strict release
protocols. "When a project approaches
EOL, maintainers are guided through
a formal termination review process.
This includes communication to the
community, archiving of resources and
decommissioning of infrastructure under
the foundation's management, drastically
reducing opportunities for hijacking or
impersonation.
"Moreover, Eclipse Foundation's neutral IP
and legal framework keeps project assets
separate from individual or corporate
identities. This protects users and contributors,
ensuring that even dormant projects
remain secure under institutional control."
16
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
ACCORDING TO JAMF 2024:
Security
Trends Report
39 % of
organisations
had at least one device
with known vulnerabilities
40 % of
mobile users
were running a device
with known vulnerabilities
9 % of
users fell for
a phishing attack
Manage and Secure
Apple at work
With Jamf Trusted Access, you ensure
that only authorised users, on enrolled
devices that are secure and compliant,
can access sensitive data.
REQUEST
Y O U R
F R E E
T R I A L
TODAY
www.jamf.com
quantum
QUANTUM QUAKES
THE INCREDIBLE PROCESSING POWER AND SPEED OF QUANTUM COMPUTERS SIGNIFICANTLY
THREATEN TRADITIONAL ENCRYPTION METHODS. HOW DO WE GUARD AGAINST THIS?
The 'quantum threat' may not be widely
understood as yet, but it has caused
tremors across many business communities
as they grapple with what might be in
store in a post-quantum world. And what
time scale are we looking at before we reach
that point - or is it just notional and some
way down the road still?
Quantum technologies are no longer simply
theoretical, points out Samantha Mabey,
director, Digital Security Solutions at Entrust -
they're fast becoming a reality. "Specifically,
quantum computing poses a particularly
urgent challenge for cybersecurity. Unlike past
technological disruptions, we have no clear
timeline for the arrival of scaled quantum
computing. When it does arrive, and if we're
not ready, it could instantly render today's
encryption obsolete, exposing sensitive data
worldwide. Even the much-feared Y2K had
a fixed deadline. 'Y2Q', by contrast, will arrive
without warning, and it will change everything."
The good news, she adds, is that we can
prepare for the future threat of quantum
tech today, through post-quantum cryptography.
"In fact, 2025 is shaping up to be
a landmark year, as organisations and
legislative bodies alike begin laying down the
foundations for quantum-safe infrastructure
and regulators start pushing for standards
that anticipate the quantum future.
"Importantly, preparing for quantum
computing is not just about future-proofing;
it's about addressing current threats, like
'harvest now, decrypt later'. This is where bad
actors steal encrypted data now, with the aim
of unlocking it once quantum computers are
viable. In some cases, breaches may have
already occurred and organisations simply
don't know it yet. Adopting quantum-safe
standards today and ensuring that your
organisation has visibility into its
cryptographic data landscape is the best
defence we have against that future."
'UNBREAKABLE' ENCRYPTION
Palo Alto Networks points to how quantum
security uses the principles of quantum
mechanics to improve the safety of
information systems and communication
networks. "It leverages unique qualities like
superposition, entanglement and uncertainty
in quantum mechanics to develop strong
security measures that can withstand
traditional and quantum attacks. With
advancements in quantum computing,
many existing cryptographic systems, such as
RSA and ECC, are at risk of being broken."
Quantum security, says the multi-national
cybersecurity company, addresses this
challenge by:
Protecting sensitive data from
future quantum-enabled attacks
Securing communications in
quantum-safe ways
Preparing organisations for the
quantum era, ensuring long-term
confidentiality and integrity.
"Quantum security leverages quantum
phenomena like superposition and
entanglement to create unbreakable
encryption methods. Unlike classical
cryptography, which relies on complex
algorithms, quantum security uses
fundamental aspects of quantum physics,
potentially creating communication channels
that are immune to eavesdropping.
"For instance, quantum key distribution
(QKD) allows secure communication by
sharing encryption keys, where any attempt
to intercept them disturbs the quantum
states, alerting the parties involved. This
innovative approach is set to revolutionise
data protection and facilitate secure interactions,
as the world increasingly adopts
quantum computing technologies."
Traditional cryptographic systems rely on the
difficulty of solving specific math problems,
like factoring large numbers, to keep data
safe, explains the company. "Quantum
mechanics studies how tiny particles behave
on a microscopic level and uses these tiny
particles and waves to perform calculations
faster than traditional computing. Quantum
security systems enable users to know immediately
if their data has been compromsed,
18
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
quantum
thanks to the laws of quantum superposition
and entanglement. These systems
use the rules of quantum mechanics to
achieve much faster processing speeds than
today's best supercomputers."
CLASSICAL VS QUANTUM COMPUTERS
Classical computers operate on binary logic.
"Every system, irrespective of its processing
strength, utilises bits-binary digits represented
by 1s and 0s-as the fundamental information
units. These bits, embodying a dichotomy
of 'true/false' or 'on/off', are the
building blocks for all data in classical
computing."
Quantum computers transcend binary
limitations, continues Palo Alto Networks.
They leverage 'qubits', which embody both
true and false states. "This quantum advantage
allows for processing information at
a pace unattainable for even the most
potent classical systems. Classical machines
require two bits for the exact computation
quantum systems can perform with a single
qubit."
The incredible processing power and speed
of quantum computers significantly threaten
traditional encryption methods. "Quantum
computing's sheer force can compromise
public key infrastructure (PKI) and uncover
significant weaknesses in current security
systems. Quantum computing poses a
threat to cybersecurity through its potential
to break the cryptographic algorithms
that currently protect sensitive data,
communications and digital transactions."
Traditional encryption methods rely on
the computational difficulty of specific
mathematical problems. "For example,
RSA encryption, a widely used form of
public-key cryptography, depends on the
challenge of factoring large prime numbers.
Quantum computers could solve these
problems much more quickly, with their
advanced capabilities, than classical
computers."
QUANTUM MECHANICS BASICS
At the heart of quantum security lies
the foundation of quantum mechanics,
a branch of physics that explores the
peculiar behaviours of energy and
particles at the microscopic scale.
"Quantum mechanics presents ideas that
challenge classical physics. These include
superposition, where particles can exist
in multiple states simultaneously, and
entanglement, in which particles are
so interconnected that the state of one
affects the other instantly, regardless of
the distance between them."
These principles, comments Palo Alto
Networks, are crucial, as they enable
quantum systems to achieve exceptional
levels of security. "For example, measuring
a quantum system inevitably disturbs it,
making any eavesdropping noticeable -
this is vital for quantum key distribution
(QKD). Grasping these fundamental concepts
of quantum mechanics is important,
because they underpin innovative
security protocols that aim to address
new cybersecurity challenges in our
increasingly digital environment."
QUANTUM KEY DISTRIBUTION (QKD)
Quantum Key Distribution relies on
quantum mechanics for what the
company claims is exceptional security.
"Unlike traditional cryptographic methods
that use complex mathematical algorithms,
QKD utilises the unique properties
of quantum particles, like photons, to
exchange cryptographic keys securely.
Keys are encoded in quantum states,
which are highly sensitive to external
observations.
If an eavesdropper tries to intercept
the transmission, the quantum state
is disturbed, instantly notifying the
communicating parties of a security
breach. Thus, QKD detects unauthorised
access and prevents successful key acquisition,
making it resistant to threats from
Samantha Mabey, Entrust: scaled
quantum computing could instantly
render today's encryption obsolete.
quantum computing."To prepare for this
future, organisations should adopt defence-indepth
strategies, considering data protection
in transit and at rest, and to be agile in the
face of emerging threats," says James Dargan,
a writer and researcher at The Quantum
Insider. "This includes network segmentation,
leveraging 5G private networks, Zero Trust
architectures and re-encrypting old files with
new technologies. The approach aims to
provide comprehensive coverage against
various attacks, preparing for the quantum
computing era while handling current
cybersecurity challenges."
As the quantum computing landscape
evolves, staying informed on the latest
developments and their implications for
cybersecurity is crucial, he adds. "By understanding
the advancements in quantum
computing, organisations can better prepare
for potential threats and implement robust
security measures. Focusing on agility and
a proactive stance in cybersecurity strategies
will ensure resilience against both current
and future challenges."
www.computingsecurity.co.uk @CSMagAndAwards July/August 2025 computing security
19
health
HEALTH ISSUES
THE NHS IS A HUGE AND VULNERABLE TARGET FOR CYBER AND RANSOMWARE ATTACKS.
WILL THE RECENT EXTRA CASH INJECTION BE ENOUGH TO KEEP IT SAFE?
Spencer Starkey, SonicWall: the
ramifications of an attack on the
healthcare sector can be disastrous.
The NHS has received a record cash
investment in the latest UK Spending
Review, with a 10% technology budget
increase. Indeed, the Government has
prioritised the NHS in the UK review, in
recognition of the fact that the NHS has
become one of the most vulnerable critical
infrastructures.
The healthcare sector, dealing with sensitive
personal information and running largely on
legacy systems, is uniquely attractive and
vulnerable to cyber-attackers. And the wave
of attacks it has suffered has clearly created
a deep malaise. Reports reveal that only 36%
of NHS staff believe cyber security measures
are sufficient.
The independent BT online survey of 76 NHS
staff at 59 NHS organisations and integrated
care systems, carried out in September last
year, explored sentiment around digital
healthcare in the UK. It found that only
42% of NHS staff surveyed trust that existing
systems are robust enough to safeguard
sensitive patient data and 64% report that
patient data is isolated and inoperable due
to outdated systems.
Despite a rise in training on new technologies
from 5% in BT's 2022 survey to 15%
in the 2024 survey, training on both new and
existing systems has fallen from 47% to 39%,
with 60% of frontline staff surveyed calling
for more cyber security training.
PRIME TARGET
SonicWall data reflects this, confirms the
company, showing healthcare remains a
prime target with hospitals and healthcare
systems taking 60-150 days to patch vulnerabilities,
while hackers exploit them in just
two days, creating a critical security gap.
Spencer Starkey, executive VP EMEA at
cybersecurity company SonicWall, homes
in on how threat actors target the most
important sectors of our society: "Despite
Rachel Reeves' NHS budget increase, the
healthcare sector remains a prime target
for ransomware attacks, as healthcare
organisations often have critical data that
they cannot afford to lose. If they haven't
already, healthcare organisations need
to have a plan in place to respond to
ransomware attacks and to minimise the
impact of these attacks.
EXPOSED DATA
"The healthcare sector continues to be a
prime target for malicious actors as evidenced
by the recent attacks on the NHS. Not only
do these attacks risk the potential for exposed
patient data, but any significant IT issue that
halts patient care poses an immediate threat
to life," states Starkey. "The ramifications of
an attack on the healthcare sector can be
disastrous, and it's important to place the
utmost amount of time, money and efforts
on securing them."
In recognition of the growing threats, the
NHS has ramped up its cybersecurity efforts.
A notable development is the announcement
of a £4.2 million investment by NHS England,
aimed at enhancing cybersecurity across the
health service. Key initiatives include the NHS
Secure Boundary project, a centrally managed
security service designed to protect NHS
networks from internet-borne threats.
20
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
retail
RETAIL AND RETALIATION
THE RECENT WAVE OF ATTACKS ON RETAIL COMPANIES HAS SHOWN THAT NO ORGANISATION,
WHATEVER THEIR STATUS MIGHT BE, IS SAFE FROM THOSE WHO TARGET THEIR OPERATIONS
The recent and many attacks on major
retail companies - such as M&S, Harrods,
the Co-op, Adidas and Victoria's Secret -
have all served to demonstrate how vulnerable
the sector is. The Co-op faced significant
disruptions, including empty shelves and
compromised member data, while Harrods
had to restrict internet access to prevent
further intrusion.
These incidents highlight how even the
most well-known and well-resourced retailers
remain prime targets for cyber-attacks. The
varied responses by M&S, Co-op, Adidas and
Harrods also show the importance of early
detection, containment protocols and clear
communication in managing a breach.
On July 10, 2025, the UK's National Crime
Agency (NCA) announced the arrest of four
individuals in connection with the cyberattacks
that disrupted operations at Marks & Spencer
(M&S), Co-op, and Harrods earlier this year.
The suspects-a 20-year-old woman from
Staffordshire, two 19-year-old men (one
British, one Latvian) from London and the
West Midlands, and a 17-year-old British male
were detained at their residences. Authorities
seized electronic devices for forensic analysis.
These arrests are linked to the hacker group
known as Scattered Spider, notorious for
employing sophisticated social engineering
tactics, SIM swapping and phishing techniques
to infiltrate organisations. In the case
of M&S, the attackers deployed ransomware,
leading to a six-week shutdown of online
clothing sales and an estimated £300 million
loss in operating profit.
DECISIVE ACTION
"In response to the series of recent cyber
incidents, retail businesses must act decisively,"
insists Robert Cottrill, technology director at
ANS. "Working with a cyber response team is
essential to assess the breach's full impact,
plug any vulnerabilities and restore systems
quickly. Clear, consistent communication with
customers and stakeholders will be vital to
rebuild trust."
Describing the incidents as a wake-up call
for all businesses, Cottrill says cyber security
must be a board-level priority. "Robust incident
response plans, proactive security reviews and
close collaboration with experts are critical.
In the fast-moving retail environment,
safeguarding customer data isn't just good
practice - it's fundamental to brand survival."
What has become clear is that no one and
no organisation is safe from such attacks,
whatever their status. The International
Criminal Court, for instance, was recently
targeted by a "sophisticated" cyberattack and
is taking measures to limit any damage, the
global tribunal has announced. The ICC, which
also was hit by a cyberattack in 2023, said
the latest incident had been contained, but
did not elaborate further on the impact or
possible motive.
Points out The Independent newspaper:
"The ICC has a number of high-profile
investigations and preliminary inquiries
underway in nations around the world and
has in the past been the target of espionage.
In 2022, a Dutch intelligence agency said it
had foiled a plot by a Russian spy using a false
Brazilian identity to work as an intern at the
court, which is investigating allegations of
Russian war crimes in Ukraine and has issued
a war crimes arrest warrant for President
Vladimir Putin, accusing him of personal
responsibility for the abductions of children
from Ukraine."
Robert Cottrill, ANS: working with
a cyber response team is essential to
assess the breach's full impact.
www.computingsecurity.co.uk @CSMagAndAwards July/August 2025 computing security
21
compliance
SEVEN-YEAR ITCH
COMPLIANCE IS NO LONGER JUST A LEGAL REQUIREMENT - IT'S A STRATEGIC DIFFERENTIATOR, ARGUES ONE
INDUSTRY VOICE. HOW WELL WILL GDPR COPE WITH THE GROWING PLETHORA OF NEW CHALLENGES IT FACES?
It's been more than seven years now since
the General Data Protection Regulation
(GDPR) came into force: on 25 May 2018.
To date, around 167 countries have introduced
their own data protection laws. Others
have legislation in the pipeline. Many of these
frameworks echo the GDPR's core principles
(transparency, accountability and subject data
rights), cementing its reputation as the global
benchmark for privacy.
But seven years on, it is clear we are entering
a new chapter of data protection. With the
rapid acceleration of AI technologies and the
increasing use of big data to train models
and automate decisions, can the GDPR keep
pace? What have we learned since the GDPR's
inception?
"Seven years since GDPR reshaped the data
landscape, the regulation remains a defining
force in how businesses handle personal data,
manage risk and build customer trust," says
Sean Tilley, senior director sales of EMEA at
11:11 Systems. "While the legislation may
no longer be new, its demands continue to
evolve - as do the expectations of regulators,
customers and stakeholders. Today, compliance
is no longer just a legal requirement:
it's a strategic differentiator."
According to the UK Information
Commissioner's Office (ICO), 'Data protection
by design and by default' remains a core
principle of GDPR, reinforcing the need to
embed compliance into systems from the
outset. At 11:11 Systems in the UK, we've
seen first-hand how organisations that treat
privacy and resilience as foundational
principles are better equipped to stay agile
and secure in a changing landscape."
The conversation has matured beyond tickbox
compliance. Modern businesses are
asking: how do we ensure our systems adapt,
scale, and remain secure, while supporting
innovation? Research from Gartner shows
that, by 2026, more than 70% of organisations
will treat privacy as a competitive
advantage, not just a compliance task. "Cloud
platforms with built-in security, automated
compliance reporting and resilience at their
core are now essential," adds Tilley. "11:11
Systems' integrated cloud, connectivity, and
security services help organisations respond
to evolving standards without compromising
agility or creativity.
Operational resilience is also under increasing
scrutiny. The UK's Financial Conduct
Authority and Prudential Regulation Authority
have issued guidance underscoring the
importance of 'impact tolerances' and robust
continuity planning. In today's environment
of escalating threats, disaster recovery, zerotrust
frameworks and data sovereignty are
non-negotiable."
Seven years on, GDPR has proven to be
more than a legal framework, he adds. "It's
a catalyst for resilient, secure and customercentric
operations. Now is the time to go
further - embedding compliance deeply into
the infrastructure of innovation."
NEVER-ENDING JOURNEY
Is the seven-year itch up to scratch? "In many
ways, GDPR has catalysed stronger governance,
clearer accountability and more resilient
systems, but the journey is far from over,"
responds Samantha Swift, senior director of
product strategy and marketing at Vaultree.
"Security vendors tend to forget that most
of what GDPR is about is protecting
personal data, not just securing it from
cyber criminals. Possibly the hardest pill to
swallow is that GDPR is, and has always
been, a legal text and it doesn't stipulate
a laundry list of products to go buy. In
reality, complete compliance is an illusion;
organisations are either non-compliant or
they simply haven't been caught."
Everyone still wants to talk about the big,
scary fines, she adds, but, beyond the
headlines, how much real benefit are
consumers seeing? "For most of us, the
internet is just as irritating as ever. Websites
bombard us with pop-ups and consent
banners - walls of legal jargon that no one
reads, but everyone clicks 'agree' on, just to
get them out of the way. The spirit of the
regulation was to empower users, but, in
practice, has it simply created more friction
without delivering meaningful control?"
Our data is still often misused, stolen,
held to ransom and left waving about in the
breeze, she points out. "The tech giants -
TikTok, Meta, Google and the rest - can
easily absorb the fines, although breaches
of GDPR ought to erode our trust in these
organisations; in most cases, we continue to
rely on their services, regardless. Meanwhile,
[possibly] everyone's favourite two-letter
acronym, AI, is popping up on every corner and
has the promise to revolutionise our lives.
22
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
compliance
"According to one group of researchers,
large language models are fast running out
of public human-generated data used for
training models, as soon as next year. Which
leaves us as the potential beneficiaries of
machine learning-based AI, AND the organisations
driving innovation in this space,
with a very real challenge: how do LLMs
continue to genuinely improve, without
causing a privacy (and ultimately a
compliance) disaster?
"Fortunately, breakthroughs in the
encryption space bring hope to the table:
technologies such as next-generation Fully
Homomorphic Encryption, which allows
data-in-use to remain encrypted whilst
models are trained, data is queried and
privacy-preserving data science collaboration
is achievable… without putting consumers'
personal information at risk in the way that
plain-text data or purely encryption in transit
and at rest does and without identifying the
individuals along the way."
GOLD STANDARD
"The 8th anniversary of the GDPR may fly
under the radar for many, but the birthday of
this gold-standard data protection regulation
is something worth noting," say Tiernan
Connolly, managing director in the cyber &
data resilience team at Kroll. "Most, if not all,
businesses will now have a GDPR framework
integrated into their data governance
practices, with the regulation becoming an
international standard for other nations to
follow. Although GDPR might be considered
'old news' in 2025, the ECB has committed to
renewing its focus on areas where 'persistent
sluggishness' is seen with compliance to existing
regulations [eg, BCBS239 - itself now a
10-year-old piece of regulation] in the
financial industry.
"This shows that older requirements will
not be forgotten or fly under the radar of
regulators. Hence, while compliance teams
may now be more concerned with adherence
to newer regulations, such as NIS2 [Network
and Information Security Directive], Digital
Operations Resilience Act [DORA] and the EU
AI Act, and how they apply to internal data
governance, protection and management,
the GDPR's landmark 4% fines loom in the
background for any business that forgets
the grandfather of data privacy regulation."
GREATER AGILITY
The compliance landscape is no longer static,
comments Sam Peters, chief product officer,
ISMS.online. "Standards like NIS 2, DORA
and the EU AI Act have recently joined
GDPR as regulations that companies need to
comply with. Additionally, the approach the
government is taking with the UK's Cyber
Security and Resilience Bill, which is due to
come into effect later this year, is to be more
agile when it comes to regulations - particularly
as new technologies, such as AI, and
threats emerge. As the government says,
'It is important for national security that our
regulatory framework is not stagnant'."
Despite this continually evolving landscape
and the introduction of new regulations,
compliance should not be seen as a blocker
to innovation, he adds. Compliance is a
catalyst when embedded early. "Take GDPR,
for example. It began as something companies
initially saw as yet more regulatory
compliance. But now it is shaping business
models and influencing everything from
product design to customer engagement.
For me, this highlights that we can encourage
better business practices, whilst also driving
profitability and innovation. Embedding
compliance in development cycles helps build
resilience from the ground up, and reduces
the cost and complexity of retrofitting
controls later."
But what types of systems should organisations
adopt that will enable them to
embed compliance in development cycles
and not stall innovation? "Adaptive, scalable
compliance systems allow both the compliance
team and the product team to work
in harmony from day one. Systems like
Sam Swift, Vaultree: organisations’
data is still often misused, stolen, held
to ransom and left waving about in
the breeze.
Sam Peters, ISMS.online: GDPR is
now shaping business models and
influencing everything from product
design to customer engagement.
www.computingsecurity.co.uk @CSMagAndAwards July/August 2025 computing security
23
compliance
Sean Tilly, 11:11 Systems: GDPR remains
a defining force in how many businesses
handle personal data, manage risk and
build customer trust.
Tiernan Connolly, Kroll: most businesses
will now have a GDPR framework
integrated into their data governance
practices.
ISMS.online's are built for adaptability and
scale. Supporting over 100 frameworks, our
systems enable organisations to pivot quickly
as new regulations emerge, without having
to reinvent the wheel," adds Peters.
"These pre-mapped frameworks and
reusable content enable businesses to build
once and apply the same solution multiple
times, achieving compliance faster while
reducing friction for development teams.
Similarly, these types of systems can help to
future-proof the business, supporting expansion
into new markets and industries, without
requiring the overhaul of security and privacy
practices each time.
"A good example we've seen is developers
integrating consent management, data
minimisation and automated logging of
processing activities into their platforms at
the build stage. This essentially means they're
making privacy part of the user experience,
not just seeing it as something another team
should handle in the back office."
CATALYST FOR CHANGE
"Once viewed as a compliance checklist, the
GDPR has been a catalyst for cultural and
strategic change," says the Data Protection
Officer (DPO) Centre. "Public awareness
around privacy has grown significantly, and
organisations now recognise data protection
as central to trust, brand value, and longterm
resilience. It has also set a global benchmark,
prompting countries worldwide to
rethink the scope and ambition of their own
data laws.
The big question is, of course, can the GDPR
keep pace with AI technologies and big data?
"AI systems are evolving fast and often rely
on vast datasets that challenge the GDPR's
principles around transparency, purpose
limitation and data minimisation," says the
DPO. "As models become more complex,
so do challenges for organisations operating
in the EU that need to comply with additional
regulations.
David Smith, DPO and AI sector lead,
emphasises how AI systems that process EU
personal data are still subject to the GDPR
and must also meet the requirements of the
AI Act. "There's significant overlap between
the two and strong data protection practices
often support AI compliance," he says. "It's
not a choice between complying with one
or the other - both are essential."
GDPR AND THE FUTURE
The GDPR was designed as a principlesbased
framework, intended to be flexible
and adaptable across different technologies
and use cases. But, despite this, many
privacy professionals believe it needs clearer,
more consistent guidance to remain effective,
especially as new technologies emerge
and complexity grows in highly regulated
sectors.
Lawrence Carter, DPO and life sciences
sector lead, says the main issue to date is
that the GDPR hasn't supported a model
for issuing secondary legislation to codify
complex issues. "Life Sciences organisations,
in particular, face ongoing challenges in
certain areas, including selecting the correct
lawful basis for clinical trials, pseudonymisation
and international data transfers.
Guidance from the EDPB and rulings from
the CJEU are, at times, inconsistent, contradictory
and not uniformly interpreted across
Member States, leading to a patchwork
of jurisdictional exceptions, rather than
harmonisation and regulatory certainty.
Introducing a mechanism to formally clarify
and update the GDPR could offer greater
consistency and confidence for organisations
navigating high-risk processing," Carter
suggests.
As the GDPR moves into its eighth year,
data protection is no longer just about
compliance. With AI, regulatory divergences
and growing public expectations, the organisations
that will thrive are those that treat
privacy as a strategic priority... built-in, visible
and always evolving.
24
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
attacks round-up
DATA THEFT ONSLAUGHT
ATTACKS AND BREACHES ARE SHOWING NO SIGNING OF EASING OFF - IN FACT,
THEY HAVE INTENSIFIED, WITH VAST AMOUNTS OF DATA EXPOSED AND STOLEN
Four people - a 20-year-old woman and
three teenage boys - were arrested in
early morning raids over the wave
of cyber-attacks that crippled M&S, the
Co-op and Harrods. Suspected of hacking,
blackmail and money laundering, the group
allegedly unleashed ransomware that stole
millions of customer records, shut down
online orders and left supermarket shelves
bare. M&S alone faces a £300 million hit,
with some systems offline for months,
highlighting the long term damage cyberattacks
cause.
Spencer Starkey, executive VP of EMEA
at SonicWall, points out how vital it is that
every single business has a robust roadmap
in place to deploy, if and when an attack
happens. "The preparation always begins
with prevention: layered security systems
and updated employee training are basic
principles in today's risky environment.
Everyone involved should have a welldefined
role and key responsibilities before
the crisis occurs."
In other attacks, a "significant amount" of
private data, including details of domestic
abuse victims, has been hacked from Legal
Aid's online system. The Ministry of Justice
said the agency's services were hacked in
April and data dating back to 2010 was
downloaded. The BBC understands that
more than two million pieces of information
were taken. The breach covers all
areas of the aid system - including domestic
abuse victims, those in family cases and
others facing criminal prosecution. "This
data may have included... addresses of
applicants, dates of birth, national ID
numbers, criminal history, employment
and financial data such as... debts and
payments," the MoJ revealed.
Qantas is another recent high-profile victim
whose customer data has been compromised.
The incident occurred when a cybercriminal
targeted one of its airline contact
centres and gained access to a third-party
customer-servicing platform. There is no
evidence that any personal data stolen, states
the airline, adding: "Qantas has reconfirmed
no credit card details, personal financial
information or passport details were stored in
this system and therefore have not been
accessed."
TRUST MUST BE REBUILT
In response to such incidents, retail businesses
must act decisively, says Robert Cottrill, technology
director at ANS. "Working with a cyber
response team is essential to assess the
breach's full impact, plug any vulnerabilities
and restore systems quickly. Clear, consistent
communication with customers and stakeholders
will be vital to rebuild trust. The
incidents are a wake-up call for all businesses:
cyber security must be a board-level priority.
"Robust incident response plans, proactive
security reviews and close collaboration with
experts are critical. In the fast-moving retail
environment, safeguarding customer data
isn't just good practice - it's fundamental to
brand survival."
Meanwhile, the Business Digital Index (BDI),
created by Cybernews, has evaluated the
cybersecurity postures of 75
European Union government
institutions and found that 67%
received a D or F rating - placing
them in high-risk or critical-risk
categories. The BDI also revealed
that every institution in the study
had experienced at least one data
breach.
www.computingsecurity.co.uk @CSMagAndAwards July/August 2025 computing security
25
artifical intelligence
AI - FRIEND OR FOE?
'AI MAY ACCELERATE OUR DEFENCES, BUT IT CANNOT REPLACE
THE HUMAN CRITICAL THINKING THAT UNDERPINS TRUE SECURITY'
- ADAM WINSTON, WATCHGUARD TECHNOLOGIES, PICTURED LEFT
According to recent statistics, 75% of
employees using AI to tackle everyday
tasks may lead to efficiency gains.
"Credential theft, data exfiltration and
infrastructure weaknesses are all potential
threats when AI tools operate without
governance," warns Adam Winston, field
CTO for managed services at WatchGuard
Technologies. "With an estimated 14% of
organisations having formal AI policies,
most deployments are left untracked and
potentially hazardous."
AI's transformative power is undeniable,
he acknowledges. "Over the last decade,
security teams have struggled under the
sheer volume of data, a fast-changing threat
landscape and a shortage of skilled people.
Machine learning models are now capable
of performing many of the repetitive duties,
such as pattern recognition, anomaly
detection, event correlation and alert triage.
All at a speed no human could match. In
incident response scenarios, AI can even
trigger containment measures before an
analyst is alerted."
Yet these gains come with their own set of
problems, Winston adds. "AI systems are
all about the quality of their training data.
Models fed narrow, outdated or even
fundamentally biased datasets may overlook
new threats or reproduce the bias at speed
and scale. Adversaries are quick to exploit
these cracks. "Techniques like data poisoning
and evasion attacks can skew ML, while
generative AI enables phishing campaigns so
convincing that they can bypass traditional
filters." One of the more hidden risks is that
decision making becomes opaque. "When an
algorithm flags suspicious activity without
explanation, security teams confront a black
box without a clear rationale and no audit
trail. In regulated sectors, such as finance and
healthcare, this lack of context can translate
directly into compliance failures and hefty
fines. But the answer isn't to abandon AI.
"It just needs to be used in the right way
and alongside human insight. Skilled analysts
bring the contextual awareness, ethical
judgment, and the legal and business understanding
that machines lack. They ask the
probing questions AI cannot. For example,
was that late?night login a genuine executive
working abroad or a sophisticated impersonation?
Does this network spike a malicious
or a flawed software update?"
Some managed detection and response
providers are already embracing this hybrid
model, he states, where you have automated
detection paired with expert investigation
and continuous tuning. "And with regulations
like the EU's Artificial Intelligence Act mandating
transparency, accountability and human
oversight in high-risk AI applications, businesses
will have no choice but to use this
approach. AI may accelerate our defences,
but it cannot replace the human critical
thinking that underpins true security. Success
will belong to those who treat AI as a force
multiplier, without losing sight of the human
expertise that steers it.
PRIME TARGET
"All AI relies upon Application Programming
Interfaces (APIs) to quickly and efficiently
communicate with other APIs and to retrieve
and condense data, making these APIs
a prime target for threat actors, cautions
James Sherlow, systems engineering director,
EMEA at Cequence Security. "In fact, Gartner
predicts that by 2028 a quarter of all enterprise
breaches will be traced back to AI agent
abuse from both external and malicious
actors, so protecting these APIs needs to
be a prime consideration for the CISO."
The challenge is compounded when
employees use shadow AI - unsanctioned AI
tools that bypass enterprise security controls
entirely. "When corporate data flows through
these personal AI accounts and unapproved
platforms, it creates additional API connections
outside the organisation's visibility and
governance framework."
One of the most important steps the
security team can take to protect these APIs
is to determine which have authentication
and access vulnerabilities, Sherlow continues.
"This is no mean feat, as discovering the build
components for many third-party APIs can
be tricky and the sad reality is that many of
the popular APIs in use by enterprise development
teams have little to no authentication
built in.
"Discovery is therefore vital in mitigating the
threat posed by AI in the enterprise. Looking
for hidden, deprecated and shadow APIs can
allow the team to understand which APIs are
in use or have fallen into disuse and which
could be actively exposing sensitive data. This
should be conducted on a continuous basis
and mapped to known behavioural models
to monitor and assess API call activity."
Also, APIs should be compliant, in that they
meet their design specifications, use appropriate
authentication, regardless of whether
they are active, and are patched when
26
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
artifical intelligence
necessary. "Protecting the APIs that underpin
AI is critical for the enterprise to detect and
resolve issues. The faster the business can
determine how APIs are being attacked, the
speedier it can respond. Utilising behavioural
fingerprints of threat actors and attacks, it's
possible to augment alerts with information
on the source IP address, [determine] which
APIs are involved and the types of data
resources being exposed, following which API
calls can be blocked, rate limited or deception
techniques used to funnel the attack."
It's only by addressing APIs that we can
hope to effectively secure AI and the next
evolutionary step: agentic AI. "By 2028,"
adds Sherlow, "a third of enterprise software
applications are expected to include agentic
AI, according to Gartner, which means it,
too, will become a prime target."
Lucy Finlay, director of secure behaviour and
analytics at Redflags from ThinkCyber, offers
a quote from Nobel Prize-winning economist
Paul Krugman in 1998: 'By 2005 or so, it will
become clear that the internet's impact on
the economy has been no greater than the
fax machine's'. "History has a habit of
repeating itself," she points out, "and we're
currently in the same position with AI, with
'experts' taking their stance on where the
chips will ultimately fall: is AI going to
precipitate the next industrial revolution?
Or is it the next fad like NFTs? Or something
more sinister, with mass job losses and the
disappearance of critical thinking on the
horizon? This hasn't stopped the public and
businesses flocking to this newfangled tool,
with AI being the mot-du-jour for investors
and consumers alike.
"At Think Cyber, we have seen the trends
in the usage of GenAI tools rocket since the
start of 2024 across our customers. As a
result, security, data protection and tech
teams are wrangling with where they're
going to place their bets. Do they acknowledge
that people are going to use this tool
and allow its use with guardrails implemented?
Do they shut it down and face the
wrath of onlookers who say they're stunting
innovation? The answer is, much like Paul
Krugman's famous statement, that experts
don't really know what the right answer is.
Anyone who professes to know what AI's
effect on the world has a good chance of
also saying something that will be just as
regrettable in 20 years' time."
In the meantime, wonders Finlay, what's a
pragmatic approach to dealing with this new
tool? "A quote from the author Kevin Sands
somewhat sums it up: 'It is never the tool
that decides. It's the hands-and the heart-of
the one who wields it.' Educate those who
are going to inevitably use the tool. At a local
level, nudge them towards critical thinking
around the consequences of their actions.
Steer them towards the controlled instances
of AI that have been thoroughly vetted by
trusted internal teams, rather than opensource
versions of the tool. Integrate into
their new workflow, merging positive feedback
for desired behaviour and personalising
guidance to their use case."
THE AI BATTLEGROUND
The AI landscape has become a battlefield
where most businesses are rapidly losing
ground, comments Tyler Reguly, associate
director, R&D at Fortra. "This battle is being
fought from multiple sides and not everyone
is following the rules. Infosec teams face so
many threats that it can be impossible to
determine the greatest threat. From prompt
injection attacks and shadow AI, to advanced
deepfakes and AI hallucinations, the attacks
come from all sides."
AI has caught the world by storm, Reguly
says. "It has given unskilled individuals the
perception that they can perform skilled
work. From vibe coding and art generation,
to eloquent writing and spreadsheet mastery,
people are given the ability to operate
outside their skillset. It is tempting to use AI
tools of unknown origin, share confidential
data and present the results as your own. At
James Sherlow, Cequence Security:
Gartner predicts that by 2028 a quarter
of all enterprise breaches will be traced
back to AI agent abuse.
Lucy Finlay, Redflags from ThinkCyber:
constructive discourse of safe usage
of AI is needed, with governmental
and AI innovators' input.
www.computingsecurity.co.uk @CSMagAndAwards July/August 2025 computing security
27
artifical intelligence
Tyler Reguly, Fortra: a complete AI policy
is imperative, ensuring your users know
what is, and is not, acceptable.
Ravit-Sadeh, CTERA: leadership teams
that delay AI policies out of fear of
'slowing innovation' are missing the
point.
the same time, senior leadership at companies
can't help but see savings when they can
replace a skilled worker with someone less
skilled."
AI hallucinations still plague generated
data, adds Reguly. "Unskilled individuals are
unaware of the mistakes that riddle their
results. This is why media literacy and skilled
employees are critical in the utilisation of AI
within organisations. Organisations must
ensure that they hire and retain skilled
individuals. "After you deal with AI hallucinations,
what about shadow AI, where
employees use unapproved AI platforms?
If you don't have an AI policy, it is hard to
argue that shadow AI exists. A complete
AI policy is imperative, ensuring your users
know what is, and is not, acceptable. Other
important tools, like data loss prevention and
cloud access security brokers, can help you
identify when AI services are accessed."
Finally, you have business impacts from
prompt injection attacks and advanced
deepfakes. "Consider a phishing email with
an attachment containing hidden prompts
that requests immediate AI-based processing.
Your entire internal approved AI infrastructure
may be compromised via a single
email. Additionally, we're seeing convincing
deepfake job applicants, utilising fake audio
and video, with nefarious intentions. These
situations are difficult to address. Human
risk management is crucial. By training
employees, you can help them spot phishing
emails and identify deepfakes."
In the rush to embrace AI's potential, many
companies are neglecting the basics, states
Ravit Sadeh, VP of product management,
CTERA. "Where is data going? Who's
accountable for its use? What happens when
something goes wrong? In my work, I've
spoken with many companies that have
had many horror stories. Employees aren't
reckless; they're trying to meet deadlines and
stay competitive. But without formal support
or training, sensitive data often ends up in
external tools, with little thought to where it
might go." And the risks aren't just internal.
"AI-generated phishing attacks have become
far more sophisticated, with attackers using
tools like ChatGPT to craft flawless emails and
even deepfake voice calls. Fake customer
service chatbots, powered by AI, are now
impersonating brands with alarming realism."
Leadership teams that delay AI policies out
of fear of 'slowing innovation' are missing
the point, she adds. "Innovation without
guardrails isn't agility, it's exposure. This
doesn't mean banning AI. It means recognising
that it's already here. The companies
best positioned to benefit from generative AI
are those who treat it as a critical piece of
infrastructure - not a toy, not a trend and
definitely not someone else's problem."
What should that look like? Start small:
Define which AI tools are approved,
and what data is off-limits
Educate employees about risks like
credential leaks and phishing
Create internal sandboxes where
teams can experiment safely
Appoint a cross-functional task force,
including IT, legal, HR, and security,
to update policies as tools evolve.
"These actions are simple, actionable and
urgent," says Sadeh. "AI use in the workplace
isn't 'emerging'; it's already widespread.
Pretending otherwise isn't just naïve, it's
dangerous. The bottom line: AI itself isn't
the threat. But ungoverned, invisible and
unmanaged use of it absolutely is."
Companies are integrating AI-powered
applications into their ecosystems faster
than they can secure them, cautions Neil
Roseman, CEO, Invicti. "The new tools carry
new risks and are often built or adopted
outside standard development pipelines,
bypassing traditional software security and
quality processes." When those applications
process sensitive data, interact with internal
systems and sit exposed on the internet,
that's a major security problem. Better
28
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
artifical intelligence
understood are the privacy risks from
employees revealing confidential information
to non-sanctioned platforms, usually GenAI
chatbots. "While not always malicious, it's
usually a symptom of inadequate internal
guidelines," he says. "The consequences
aren't obvious, but can be significant, if your
company secrets somehow get into an
external AI model with hazy privacy policies.
The security and privacy risks are very real,
but also known and addressable. Security
tools already exist to identify exposed AI
assets and scan them for vulnerabilities. Data
privacy concerns can be mitigated through
clear terms of use and robust internal policies
that distinguish between sanctioned and
unsanctioned AI usage."
But there's a third, far more insidious, risk
emerging: overreliance and excessive trust,
he says. "As AI-generated summaries and
insights become commonplace efficiency
boosters, it's all too easy to start trusting
machine output as a reflection of reality. And
when business decisions are made, based
on unverified AI responses, the line between
efficiency and irresponsibility begins to blur."
Dependence on the functionality and data
fidelity of AI platforms carries strategic risk.
"It's one thing to pilot an AI project and fail to
realise value. It's another to architect entire
workflows, customer interactions or product
strategies around systems that may hallucinate,
misinterpret or simply go offline. If the
model fails, who's accountable for business
downtime? If the data is wrong, who takes
responsibility for the bad decisions it
backed?" asks Roseman.
"We're already learning to address AI
security gaps and privacy concerns, but overreliance
is a different beast. The companies
that will thrive in an AI-driven economy
aren't those that adopt AI fastest or most
extensively - they're the ones that deploy it
most thoughtfully. Right now, the biggest
threat isn't AI going rogue. It's us outsourcing
our thinking to it. In the absence of proper
governance, organisations are already
suffering the consequences, comments Dave
McGrail, head of business consultancy at
Xalient. "Recently, hackers breached a popular
AI chatbot service, exposing thousands of
user chat logs, along with credentials and API
keys buried in those conversations. This kind
of data leakage can be catastrophic to a
business, in terms of potential regulatory
penalties and damage to brand reputation.
"There are a growing number of poorly
secured chatbots being used against their
owners. Meanwhile, cyber-criminals are
exploiting the AI boom with a fake 'ChatGPT'
browser plugin which stole login credentials
[more than 4 million in one haul] from users
drawn in by the AI craze. These incidents,
spanning data leakage to credential theft,
highlight the breadth of threats when AI is
adopted without oversight."
The common thread here is that
uncontrolled use magnifies security gaps.
"Organisations must proactively bring AI
usage out of the shadows, says McGrail.
"This starts with clear internal policies and
training on what data employees can feed
into AI systems and which tools are
approved. Unsanctioned AI tools should be
treated as the new shadow IT and be subject
to the same scrutiny as any unscreened
app or cloud service. Technical controls to
monitor and restrict sensitive data going into
AI queries should be implemented, and
vendor security due diligence ensured to
capture and treat risk for any AI platforms.
Fortunately, guidance is emerging, he
continues. Frameworks such as the NIST AI
Risk Management Framework and ISO/IEC
23894:2023 (AI risk management guidance)
offer blueprints for governance and controls.
"At a time when AI innovation and adoption
is outpacing most company compliance
and governance playbooks, a thoughtful,
accountable AI governance program should
be viewed as a business imperative, rather
than a burden."
Neil Roseman, Invicta: right now, the
biggest threat isn't AI going rogue. It's
us outsourcing our thinking to it.
Dave McGrail, Xalient: organisations
must proactively bring AI usage out of
the shadows.
www.computingsecurity.co.uk @CSMagAndAwards July/August 2025 computing security
29
legislation
NEW CYBER LAWS MUST BE 'MORE AMBITIOUS'
MPS SHOW WIDESPREAD SUPPORT FOR STRONGER CYBER LAWS, ALONG WITH CALLS FOR GREATER
GOVERNMENT COLLABORATION AND A MORE AMBITIOUS, FUTURE-PROOFED APPROACH
Dan Aldridge, chair of the APPG
for Cyber Innovation.
Till Sommer, Policy Counsel at
Cyber security Business Council.
The first public statement from MPs
on the forthcoming Cyber Security
and Resilience (CSR) Bill highlights
a sense of cautious optimism, with 46%
of survey respondents believing the CSR
Bill will support economic growth, while
44% merely "see the potential".
Against the backdrop of high-profile
cyberattacks on major UK retailers, the
Cyber Innovation All-Party Parliamentary
Group's official report further underscores
the urgent need for a more ambitious and
inclusive approach to cyber legislation.
The report was developed based on
insights from a national survey, with
89 respondents from across the cyber
sector and beyond. The report incorporated
inputs from a parliamentary roundtable
discussion held under the Chatham
House rule that brought together 17
representatives from managed service
providers, cyber companies, academics
and other organisations in the UK.
NARROW APPROACH
The CSR Bill presents a transformative
opportunity to update the UK's cyber
legislation, yet - according to the Cybersecurity
Business Network - currently
adopts a narrow approach, which excludes
key opportunities that would benefit the
UK economy and wider society, including:
Embedding corporate governance
at the heart of corporate decision
Empowering cyber professionals
to address the emerging threats
Providing legal protection for threat
intelligence
Aligning regulating requirements
to reduce compliance issues and
drive higher standards.
Are there failings in the Bill that need to be
addressed? Till Sommer, Policy Counsel at
Cyber security Business Council, sees the
issues differently. "The problem is less about
the Bill being weak. The Bill, at least based
on what we know about it at present, is
totally doing the right thing, but my worry
is that the Government is just a little bit too
conservative. This will be the first Bill ever
with cyber in the title and while I totally get
concerns about this becoming a Christmas
tree bill to which you attach every single
cyber issue, I am concerned that we are only
looking at a very small slice of a cake that is
made up of lots of unsolved problems when
it comes to cyber security and cyber regulation.
Getting parliamentary time to pass
legislation is incredibly difficult and we need
to make the most of this opportunity."
The Cyber Innovation All-Party Parliamentary
Group (APPG) believes there's a risk
the CSR bill will not receive the necessary
support from stakeholders, unless a more
comprehensive approach is taken to engage
with relevant stakeholders before and during
the parliamentary process.
Dan Aldridge MP, chair of the APPG for
Cyber Innovation, comments: "This Bill is a
historic opportunity to strengthen the UK's
cyber resilience, but we risk falling short,
if we don't listen to those on the frontline.
We're calling on DSIT [Department for
Science, Innovation and Technology] to
open up the conversation, coordinate across
government, to provide a timeline and
process for tackling the urgent issues that
are deemed out of scope. By future-proofing
regulations and giving Parliament a clear
role in oversight, we can make sure the
UK remains secure and competitive in a
rapidly changing digital world."
30
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
data protection
16 BILLION REASONS TO TAKE ACTION NOW!
DATA BREACHES THAT LED TO 16 BILLION PASSWORDS BEING STOLEN COULD
CREATE A SNOWBALL EFFECT OF CYBER-ATTACKS IN THE DAYS AHEAD
Xavier Sheikrojan, Signifyd: balancing
advanced technology with human
oversight will be essential in addressing
the fallout from this breach.
How big can data theft really get?
Extremely so, is the answer, particularly
in the wake of the discovery of
one of the largest data breaches in history:
several collections of almost 16 billion
exposed login credentials.
The research that unearthed the theft,
based on unique Cybernews findings and
originally published on its website on
18 June, suggests the data most likely
originated from various infostealers.
Xavier Sheikrojan, senior risk intelligence
manager, Signifyd, believes the data
breaches that resulted in these 16 billion
passwords being stolen could create a
snowball effect of cyber-attacks. "A key
concern is the rise of sleeper accountsaccounts
created using stolen details that
can be used not only for immediate fraud,
mimicking legitimate customers before being
exploited at scale. Although it's too early to
draw definitive conclusions, the timing
of these increases aligns with the recent
breaches, suggesting fraudsters may
be testing the waters with stolen data,
potentially through account takeover and
credential-stuffing attacks."
He advises businesses to stay vigilant and
implement robust protective measures, such
as monitoring for anomalies in behaviour
from their existing users and customers.
"Sometimes hackers only need one set of
matching stolen credentials, so a forced
reset of passwords, using strong and
unique passwords, and using two-factor
authentication, can be great strategies.
This not only protects the business, but
also safeguards loyal customers. If you
have manual review teams, ensure they
are educated and aware of the latest data
breach trends. Additionally, proactively find
ways to optimise your machine learning
detection. Balancing advanced technology
with human oversight will be essential in
addressing the fallout from this breach."
ROTATION SALVATION
Spencer Young, SVP EMEA, Delinea, says the
current wave of data breaches shows that
static credentials can be a serious liability in
today's fast-moving world. "Passwords alone
- especially unrotated ones - leave consumers
and organisations vulnerable to phishing, credential
stuffing and Pass-the-Hash attacks.
Good password hygiene isn't enough
anymore. Credential vaulting and automated
password rotation are foundational
to stopping lateral movement. By continuously
rotating credentials and limiting their
lifespan, organisations and consumers can
invalidate stolen hashes and prevent
attackers from moving freely."
Passwordless initiatives with the aim of
reducing the risks are becoming increasingly
more popular as well. "Technologies such as
biometrics, where biometric data remains
encrypted and safely stored in the device and
does not travel across the network, improves
the authentication process because it's based
on a factor that only the user has and does
not leave their device," adds Young. Despite
these developments, passwords are not
disappearing. They are simply moving to
the background and becoming part of an
authorisation experience, with one-time
passwords, magic links, temporary keys and
just-in-time access to stay ahead of threats."
James Shank, director of threat operations
at Expel, underscores the scale of what is at
stake equally well when he says: "If this news
frightens you, then your security program
probably has some fundamental gaps.
Let this be the fuel you need to position
yourself and your department for solving
the problem systematically, rather than
defending against the news du jour. There
will always be another breach, with even
more passwords, and emergency handling
will continue, if you don't have systematic
defences in place."
The scale of apparent negligence when it
comes to lost data is startling. When, for
32
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
data protection
instance, Apricorn recently announced
the findings from its annual Freedom of
Information (FoI) requests into device loss
and data breaches across major government
departments in 2024, the figures indicated
that device security issues remain endemic
across the public sector. Several departments
reported an increase in lost and stolen devices,
compared to the previous year, despite
attempts to address the issue.
"Across the 17 departments questioned,
more than 1,200 organisational devices were
reported lost or stolen between January and
December 2024. HM Revenue and Customs
(HMRC) alone accounted for 804 of these
losses, including 499 mobile phones," reveals
Apricorn, manufacturer of software-free,
256-bit AES XTS hardware-encrypted USB
data storage devices. "While this represents
a modest decrease compared to the 1,015
devices lost by HMRC in 2023, the number
remains troubling, given the sensitivity of the
information the department handles. A large
number of the reported phone losses were
the result of an internal audit that flagged
legacy devices replaced with newer models,
highlighting ongoing inventory management
challenges."
Other departments showed a more
worrying trend, with The House of
Commons reporting 100 devices lost or
stolen during 2024, a significant increase
from 65 devices the previous year. Similarly,
the Department for Education (DfE) saw
device losses climb from 78 in 2023 to 107
in 2024. The Department for Energy Security
and Net Zero (DESNZ) also reported a rise,
from 122 lost devices last year to 150 this
year. Meanwhile, the Department for
Science, Innovation and Technology (DSIT)
reported 113 missing devices.
"Although HMRC's numbers suggest some
improvement following internal audits, the
continued high levels of device loss across
government departments show that fundamental
issues have not been resolved," says
Jon Fielding, managing director, EMEA,
Apricorn. "Every lost or unaccounted device
carries a risk for those individuals whose data
could be exposed."
The findings also reveal the extent of
personal data breaches, with The House of
Commons disclosing 49 incidents involving
personal data during 2024, up from 41
reported the previous year. Despite these
breaches, the House of Commons has not
had to disclose any such personal data
breach to the Information Commissioner's
Office (ICO) in this period. The figure
highlights the continued vulnerability of
sensitive personal information within
Parliament and other institutions.
Worryingly, several departments that had
previously been forthcoming with breach
and incident reporting declined to respond
in full this year. The Ministry of Justice (MoJ)
and the Department for Education (DfE), for
example, both refused to disclose details on
data breaches and reports made to the ICO,
citing exemptions under Section 24(2) of the
Freedom of Information Act (FOIA). The
exemption states that there is no duty to
confirm or deny whether the requested
information is held, if doing so would
prejudice national security. Several other
departments also failed to provide the
relevant statistics.
Fielding believes this growing lack of
transparency raises further questions about
the true scale of data breaches occurring
within government departments and the
threat to data. "Whilst all departments
confirmed their devices are encrypted,
they must be supported by strong back-up
protocols, inventory control and employee
awareness programmes.
"A holistic approach to data protection,
including frequent audits, multiple back-up
copies and rigorous disaster recovery testing,
is essential to minimise the risks posed by
device loss and theft."
James Shank, Expel: if this news frightens
you, then your security program probably
has some fundamental gaps.
Jon Fielding, Apricorn: the continued high
levels of device loss across government
departments show fundamental issues
have not been resolved.
www.computingsecurity.co.uk @CSMagAndAwards July/August 2025 computing security
33
quantum
QUANTUM BLAST-OFF
BATTLE TO BEAT THE HACKERS WILL BE FOUGHT IN OUTER SPACE
Researchers in Europe and Canada
are developing technology that will
allow them to fire quantum-coded
messages across continents via satellites
in space.
According to DigiKey, the project is
helping to lay the foundation for
a new kind of internet - one that is
fundamentally unhackable. Called
HYPERSPACE, it is pushing beyond the
limits of fibre-based quantum links and
using space to provide ultra-secure
communication.
"While a fully operational transatlantic
quantum link remains years away,
HYPERSPACE aims to tackle the core
scientific and technological challenges
that would make such a breakthrough
possible," says DigiKey. "The project is
widely seen as a modern echo of Marconi's
pioneering transatlantic radio transmission
in 1901, and HYPERSPACE could mark
the beginning of a quantum-powered
internet built not on cables and code,
but on entangled particles and the laws
of physics."
TOTALLY SECURE ENCRYPTION
States project coordinator Prof Dr Fabian
Steinlechner comments: "HYPERSPACE is
working on a way to generate totally
secure encryption keys at a distance
through space using quantum technology.
One day, this could connect entire
continents with communication that's
impossible to hack. Today, Europe and
Canada are building the foundation for
that future by testing how we can transmit
quantum signals between satellites and
the ground."
At the heart of HYPERSPACE is a phenomenon
called entanglement, described
by Einstein as 'spooky action at a distance'.
In a simple analogy, entangled particles
behave like identical twins. The entangled
twins can have one of two eye colours,
blue or brown, states DigiKey. "When
the eye colour of one of the twins is
determined, the eye colour of its twin is
known instantly. According to quantum
theory, however, the eye colour is not
determined until we actually measure
(or 'look at') it."
Quantum communication dispenses
with complex passwords or encryption
algorithms, but is protected by the use of
entangled particles. These quantum signals
can be sent through fibre-optic cables or
even beamed through open space, but
unlike conventional data, any attempt to
intercept them immediately breaks the
connection and exposes the eavesdropper.
Currently, most quantum communication
systems rely on photons travelling through
fibre-optic cables to share encryption keys.
But fibre-based systems on the ground can
only go so far: after a few hundred kilometres,
the signal weakens and becomes
unreliable. This is why the HYPERSPACE
team is looking to space: to explore how
quantum signals can be sent between
satellites and ground stations, enabling
secure communication over vast distances.
PHOTON FINISH
The team is exploring how to encode
multiple quantum bits (qubits) onto a
single photon to create 'high-dimensional
entanglement', essentially packing more
information at once.
"Quantum entanglement is powerful,
but usually it only enables one bit of
information at a time to be sent, like one
car driving down a single-track road,"
explains DigiKey. "With high-dimensional
entanglement, it's like adding extra lanes
to the road. Suddenly, multiple vehicles
can move side by side, which means more
information travels faster, with less chance
of slowdowns or interference."
34
computing security July/August 2025 @CSMagAndAwards www.computingsecurity.co.uk
Computing
Security
Secure systems, secure data, secure people, secure business
Product Review Service
VENDORS – HAS YOUR SOLUTION BEEN
REVIEWED BY COMPUTING SECURITY YET?
The Computing Security review service has been praised by vendors and
readers alike. Each solution is tested by an independent expert whose findings
are published in the magazine along with a photo or screenshot.
Hardware, software and services can all be reviewed.
Many vendors organise a review to coincide with a new launch. However,
please don’t feel that the service is reserved exclusively for new solutions.
A review can also be a good way of introducing an established solution to
a new audience. Are the readers of Computing Security as familiar with
your solution(s) as you would like them to be?
Contact Edward O’Connor on 01689 616000 or email
edward.oconnor@btc.co.uk to make it happen.
Computing
Security
Secure systems, secure data, secure people, secure business
e-newsletter
Are you receiving the Computing Security
monthly e-newsletter?
Computing Security always aims to help its readers as much as possible to do
their increasingly demanding jobs. With this in mind, we've now launched a
Computing Security e-newsletter which is produced every month and is available
free of charge. This will enable us to provide you with more content, more
frequently than ever before.
If you are not already receiving this please send your request to
christina.willis@btc.co.uk and advise her of the best email address for the
newsletter to be sent to.