Kenya's president is pushing for the UN to take a leading role in global AI governance—a position that reflects how the Global South views the current fragmentation of regulatory efforts. Most AI policy-making has concentrated among wealthy nations and tech companies, leaving developing…

Ruto urges world to centre the United Nations in AI governance

In July this year, Kenya launched a nationwide public participation exercise on its draft AI policy.

nation.africa

Most people encounter far less misinformation than the narratives of an information apocalypse suggest—but the distribution is deeply unequal, and the stuff that does spread tends to cluster among already-vulnerable populations. This study cuts through the hype by measuring actual exposure…

European security agencies are grappling with a shifting threat landscape where the deadliest attacks often come from the hardest to stop: low-complexity operations by isolated individuals. Meanwhile, state-backed hybrid tactics and the radicalization of minors are accelerating in ways that…

China's spy agency has gone public with AI concerns for the first time—a notable shift that signals how seriously Beijing views the technology's national security implications. The Ministry of State Security's warning reflects growing alarm over AI's role in surveillance, espionage, and…

Researchers analyzed 134,000 posts from German and Dutch elections and found AI-generated imagery clustered among smaller and far-right parties—raising questions about who's adopting synthetic media and whether it actually moves voters. The striking part: emotional impact came from *what the…

UseEffectsVisualGenAICampaigns | Simon Kruschinski

What do we know from empirical studies about the use and effects of AI-generated images and videos in election campaigns? Fabio Votta and I dug into our data of roughly 134,000 Instagram and Facebook ...

linkedin.com

Swedish courts just proved that geography doesn't shield online predators: an 18-year-old has been sentenced for attempted murder and rape committed against victims in Australia and Germany while still a minor. What's significant here isn't just the conviction—it's the precedent. Most…

Swedish teenager jailed for attempted murder, rape and assault committed online, with victims in Germany and Australia

Eighteen-year-old known as Chai was active in networks notorious for their sadistic exploitation of young people

theguardian.com

UK police databases hosted on US cloud infrastructure are exposed to warrantless US government access under the Cloud Act—and that legal asymmetry creates serious problems. The GDPR assumes data protection remains under the host nation's control, but US law lets federal agencies demand data…

Sensitive UK police data vulnerable to ‘compromise’ by US government and foreign actors

Exclusive: Official UK security assessment found Microsoft cloud platform storing files was at potential risk from hostile hackers

theguardian.com

Law enforcement in England and Wales still treats online abuse as a secondary concern compared to physical harm—a gap that becomes harder to justify as technology-facilitated attacks grow more coordinated and damaging. This practitioner assessment reveals a deeper problem: most police forces…

Protecting Women Online at The Open University

Empower beginners to engage deeply with musical activities. Provide new tools and capabilities for expert musicians and theorists. Cast new light on how music and music perception work.

cpwo.open.ac.uk

Europe's new AI strategy stakes a claim on something crucial: supply chain resilience and assurance leadership. Rather than chase raw capability, the proposal builds toward state coordination on hardware security, component sourcing, and verification standards—the unglamorous infrastructure…

A Transformative AI Strategy for Europe

A newly published strategy for Europe advocating for a state alliance for supply chain security, and becoming a leader in AI assurance among other objectives.

transformative-ai.eu

Extremists have found an unexpected distribution channel in CapCut: the video editing app's remix-friendly design and young user base make it ideal for producing propaganda that doesn't carry obvious IS branding. Once created there, these videos slip into social feeds with less friction than…

“Create, Share, and Support”: How Pro-Islamic State (Is) Supporters Are Exploiting Capcut to Produce and Disseminate ‘No-Brand’ Propaganda Content Online - GNET

CapCut, with a young audience, serves both as a lab for creating new non-branded material and as a repository for existing media to remix. The resulting non-IS-branded content can be harder to detect and delete from social platforms.

gnet-research.org

Middle powers are chasing sovereign AI models—but they may be looking at the wrong prize. The real leverage lies in what sits around the model: the infrastructure that feeds data, manages permissions, gives agents memory, and ultimately shapes how AI gets deployed in the real world. That…

Researchers tested what happens when AI systems quietly steer people toward worse choices—and found people don't just accept the bad advice, they actually rewire their own preferences to justify it. Subjects rated options lower after an AI nudged them that direction, then kept trusting the…

Human preferences are susceptible to covertly misaligned AI advice | PNAS

AI assistants are increasingly used as advisors to guide decisions, yet little is known about how people evaluate such advice when the advisor’s un...

pnas.org

Curtis Yarvin's latest argument against AI existential risk has gained traction in policy circles—he dismisses doom scenarios as fantasy while insisting the real concern is AI's actual capabilities and deployment, not hypothetical superintelligence. Understanding why this framing matters for…

How the USG should respond to AI

“AI has not created new risks. It has made old risks urgent.”

graymirror.substack.com

UK adults are increasingly getting news through social media, search, and AI tools—but Ofcom's latest data shows something reassuring: these platforms aren't displacing traditional sources. Nearly 75% use online intermediaries, yet only 5% rely on them exclusively. Most people mix…

New report from Tony Blair Institute argues for the construction of an AI assurance ecosystem to 1) Deliver Better, Safer and More Trustworthy AI Products, 2) Build Justified Trust to Drive AI Adoption and Strengthen AI Markets, 3) Streamline Compliance and Facilitate Access to Global Markets,…

Real-time voice synthesis at the hardware level fundamentally reshapes the deepfake threat model. When voice modification happens before audio even reaches recording software, the "hardware-to-software chain of custody" approach to authentication collapses—there's no technical way to…

Voicemod’s new pocket-friendly device brings real-time voice changing to your phone

Voicemod’s real-time voice changing comes to iOS and Android.

theverge.com

How the Cutecore aesthetic is repurposed as a gendered mechanism of grooming, coercive control and recruitment within nihilistic violent extremist (NVE) online spaces.

Sh/Ed culture and trophies: how Cutecore and pastel aesthetics are used to recruit youths in the com network

This article examines how the Cutecore aesthetic is repurposed as a gendered mechanism of grooming, coercive control and recruitment within nihilistic violent extremist (NVE) online spaces. Cutecor...

tandfonline.com

The framing of AI risk as an existential threat can itself be a strategic move. Some argue that worst-case scenarios—whether plausible or inflated—serve industry interests by justifying regulatory capture, concentrating power among well-funded incumbents, and raising barriers to entry for…

US Tech Firms Seek to Control AI Regulation with Misinformation Campaign | Jack Perschke posted on the topic | LinkedIn

I’ve been trying to stay quiet but this “AI is going to kill us all” stuff is getting so simultaneously urgent and ridiculous that I’ve decided I have to speak-out. We are all being subjected to an ...

linkedin.com

The U.S. military has openly acknowledged development of space-based weapons systems—a significant shift from decades of ambiguity. What makes this disclosure notable isn't just the capability itself, but the strategic messaging: openly declaring these systems changes deterrence calculations,…

US military reveals it has weapons in space

Security experts warn announcement could result in the proliferation of weapons by other governments

ft.com

This review of 24 widely circulated public predictions of the future of AI shows they are often flawed. Before allowing such predictions to shape action, we should ask how reliable they are.

The debate over AI regulation keeps circling back to human rights—but what does that actually mean in practice? This piece examines how rights frameworks could (or should) shape policy, and where the tension lies between protecting people and enabling innovation. Worth reading to understand what…

State actors aren't just spreading propaganda anymore—they're actively cultivating networks that radicalize toward violence, with Russian operations now documented feeding into nihilistic extremism across platforms. The distinction matters: this isn't passive information warfare but deliberate…

‘United by chaos:’ Examining the Russian nexus to online networks of nihilistic violence - Institute for Strategic Dialogue

An ‘Alliance’ is uniting nihilistic violence with Russian-speaking groups, incentivizing offline violence. But why? ISD investigates its origins.

isdglobal.org

Russian-linked tracking pixels are harvesting data from Romanian users—a finding that illustrates how ad tech infrastructure becomes a vector for state-level surveillance when oversight breaks down. The mechanics are deceptively simple: invisible code embedded in ads collects behavioral data at…

There's a bit of me that wonders if China is now joining the AI Race narrative with the same goal as the US in the space race with Russia...flexing their economic strength and stressing the more fragile adversary with exorbitant expenditure and risk with limited returns.

The latest AI doomsayer is China’s intelligence boss

Beijing’s response is to ‘firmly grasp technological sovereignty’ and broad regulations

theregister.com

Over 100 female politicians across Europe have been targeted by deepfake pornography sites—a coordinated attack spanning 22 countries. The scale is staggering: researchers found 160 dedicated websites hosting synthetic intimate images, almost exclusively of women in office. As deepfake tools becom…

Sexually Explicit Deepfake Sites Target 100-Plus Politicians in Europe

An analysis of 160 deepfake websites reveals politicians in 22 countries appear on them. Nearly all of them are women.

wired.com

Law enforcement operates on incremental gains while criminal adaptation accelerates and that gap creates the 'interdiction paradox'..

The Future of Crime Revisited: Artificial Intelligence and the Next Phase of Crime Exponentiation

In earlier work, Nikolovska and I (2020) argued that technological change had outpaced the historical equilibrium between crime and the institutional response to it, producing a condition we termed cr...

crimrxiv.com

Crypto platforms have hemorrhaged $3.63 billion since January—a staggering figure driven by supply-chain attacks, smart contract exploits, and compromised private keys. The 2026 report maps where these breaches are actually happening, which matters because the industry's security posture shapes ri…

2026 State of Crypto Security Report | CoinGecko

Crypto lost $3.63B to hacks since 2025. See how attack vectors, actors, and defenses are evolving, and how CEXes protect user assets.

coingecko.com

Deepfake audio is moving from disinformation labs to active military operations. Houthis reportedly deployed a voice clone of an opposing commander ordering retreat—a tactic that blurs the line between information warfare and kinetic conflict. The target group has now issued guidance that only off…

"𝗔𝗹𝗹 𝘁𝗿𝗼𝗼𝗽𝘀 𝗿𝗲𝘁𝗿𝗲𝗮𝘁!" 🔊 Houthis allegedly used an 𝗔𝗜 𝘃𝗼𝗶𝗰𝗲 𝗰𝗹𝗼𝗻𝗲 of an enemy commander telling his troops to retreat. Possibly the most cyberpunk thing I've seen. The… | Marijn Markus

"𝗔𝗹𝗹 𝘁𝗿𝗼𝗼𝗽𝘀 𝗿𝗲𝘁𝗿𝗲𝗮𝘁!" 🔊 Houthis allegedly used an 𝗔𝗜 𝘃𝗼𝗶𝗰𝗲 𝗰𝗹𝗼𝗻𝗲 of an enemy commander telling his troops to retreat. Possibly the most cyberpunk thing I've seen. The AI-generated audio recording wa...

linkedin.com

Police interviews are high-stakes moments where memory, stress, and power imbalances already distort what gets said. Adding AI to that environment—whether for real-time guidance or analysis—introduces new risks around suggestion, bias, and accountability. May breaks down why the convenience of aut…

Investigative interviews are key to solving crimes – should AI be helping police with their inquiries?

The way a police interview is conducted affects the credibility of everything that follows in the criminal justice process.

theconversation.com

UK plans for media accountability on social platforms are drawing fire from the Trump administration. The tension reveals deeper questions about whose interests shape platform governance—and whether promoting trustworthy information must mean surrendering editorial independence to foreign pressure.

Trump administration attacks UK plans to boost traditional media on social platforms

US embassy says Downing Street’s plan to give BBC, ITV and other established media more online visibility carries risk of ‘censorship’

theguardian.com