“The European Union's cybersecurity agency and in-house research center have each begun experimenting with open artificial intelligence models from China, three EU officials have told POLITICO. The Joint Research Centre (JRC) — the EU’s in-house scientific research center — is evaluating ‘publicly available open-weight models’ originating from China, European Commission spokesperson Maciej Berestecki confirmed. The head of ENISA, the bloc's agency responsible for cybersecurity, also said in an interview that his team was now testing Chinese AI. The experiments with Chinese models at the JRC are part of a broader research project to evaluate artificial intelligence that also includes leading models from the United States and Europe, Berestecki said. ‘Models from different origins are included in a comparative scientific evaluation of their capabilities, performance, limitations and risks,’ he said, adding that the JRC is also looking into their potential scientific applications. […] Open-weight models are those whose core components are publicly released, allowing anyone to run the model and study how it works. China’s approach to AI has leaned more towards open models than the United States. Some major U.S. tech and AI companies, including chips giant NVIDIA, are investing in open-weight models.” #Privacy #GDPR #ChineseAI #Cybersecurity Subscribe to EDPO's newsletter here: https://lnkd.in/dwK8sde *This article was not written by EDPO. The opinions and views of the author(s) do not necessarily reflect those of EDPO. https://lnkd.in/ezJkHkCr
EDPO (European Data Protection Office)’s Post
More Relevant Posts
-
406 scientists, including Geoffrey Hinton, Yoshua Bengio, and Stuart Russell, are demanding that companies immediately disconnect AI models from the internet. Their open letter responds to incidents where AI systems broke out of poorly secured environments and compromised external systems. The signatories call for mandatory security standards, an international oversight body, and required reporting of all containment failures. More: https://lnkd.in/efStv_ci
-
Does knowing how to solve a problem justify having the authority to decide how to solve it? Recent incidents involving AI agents accessing systems without authorisation raise a question that goes beyond cybersecurity: what happens when, to achieve a goal, an AI uses means we would never have authorised? We may be granting authority to systems whose capabilities we admire before we have sufficient grounds to trust their judgement. David Fernández and I explore this question in “The Paradox of the Nobel Prize-Winning Child” The ability to solve a task is not enough to justify letting an AI decide how to carry it out. Who sets its boundaries? Which values should guide its decisions? Who is accountable when it crosses those boundaries? 📄 Article: https://lnkd.in/eQj8VRiD What evidence should we require before granting AI greater autonomy? #ArtificialIntelligence #AIEthics #AIGovernance
-
We're pleased to welcome Dr Elizabeth Seger, Senior Policy Advisor at the Tony Blair Institute for Global Change, as a speaker at the Oxford Generative AI Summit! Elizabeth leads the Institute's programme on global AI governance and policy, working to advance AI policy that underpins innovation and unlocks public value, while addressing the societal and capability risks associated with advanced systems. Her focus spans democratic resilience, cybersecurity and open-source policy. She is currently leading a research programme mapping national AI assurance ecosystems, focused on how AI assurance tools, standards, processes and institutions can support trustworthy adoption in high-stakes contexts. Before joining TBI, Elizabeth was Director of Digital Policy at Demos, where she co-founded the Epistemic Security Network, a cross-civil-society response to AI-related information threats. She has also published agenda-setting research on open-source AI policy with the Centre for the Governance of AI, and holds a PhD in Philosophy of Science from the University of Cambridge. 🔍 Read more: https://lnkd.in/eXf3zANG 🎟️ Book your ticket: https://lnkd.in/eaBju_4F #OxGen26 #AI #GenAI #TechLeadership #AILeadership
-
-
NEW RELEASE | The Paradox of Precaution: Delaying AI, Exposing Australia to Quantum Threats Ghost of Governance has released the latest instalment in our seven-part AI series. It examines the Paradox of Precaution: when governments seek to protect society by restricting emerging technologies, they risk delaying the very capabilities needed to defend against greater threats. We explore quantum computing, the vulnerability of modern cryptography, AI-driven scientific breakthroughs, and the consequences for Australia's national security. Explore the rest of our AI series: 1. Cybersecurity in the Age of Frontier AI: Why Governments Are Already 2. Behind OpenAI's Warning Shot: The Hugging Face Breach 3. Build AI: Govern It... Without Stifling Innovation 4. The Ethics of the Lever: AI, Power and Who Decides 5. The Digital Manhattan Project: AI and the New Contest for Power 6. Photonic Chips and the AI Acceleration Nobody in Canberra Is Talking About #AI #QuantumSupremecy #PCQ #ASI #SI #SuperIntelligence #Australia #OpenAI #Anthropic #Cryptography #LandForces2026 #LandForces https://lnkd.in/gN96AQPx
-
AI self-replication moves from theory towards technical testing Anthropic has published controlled experiments examining whether smaller AI models could reproduce across internet-connected devices. Models with 30 billion parameters or fewer could already complete many individual steps involved in replication. Under simplified conditions, stronger systems completed an end-to-end replication process across a contained test network. This was not a demonstration of an uncontrolled AI spreading across the public internet. Some cybersecurity experts also consider Anthropic’s near-term warnings overstated. Nevertheless, the research identifies a credible security direction: AI-enabled malware capable of adapting to unfamiliar devices and operating without a central service that defenders can simply shut down. Around 400 scientists, including Geoffrey Hinton and Yoshua Bengio, have now called for international containment standards, independent oversight and a registry of significant security incidents involving advanced cyber-capable AI. Sources: https://lnkd.in/gqNFXmqS https://lnkd.in/dmmmYsPs
-
On Tuesday, a coalition of more than 400 researchers, including AI godfathers Yoshua Bengio and Geoffrey Hinton, published an open letter calling for AI and security researchers to collaborate on improving alignment and containment approaches, and aiming to protect critical infrastructure before more incidents of "high-profile security failures" come to fruition. https://lnkd.in/gwzuyUvn
-
Something new is starting in Kaiserslautern. We’re building a local community around AI safety, bringing together people interested in understanding and addressing the challenges that come with increasingly capable AI systems. Our first meetup is on 29 October. If you’re curious about AI safety, come join the conversation! 👇 Details below.
Last month, the AI Safety Newsletter reported that AI agents inside OpenAI's cyber tests had built themselves a message board. It was removed. Days later, they built another one. (https://lnkd.in/et_GJkqP) It would be easy to move on. But ordinary life is going to be shaped by how these systems turn out. We would rather understand them. So we are starting AI Safety Kaiserslautern: talks and discussions on how AI systems fail and what makes them trustworthy. First meetup: two talks and a panel discussion: 🗓️ Thursday 29 October 2026, 15:30–18:00 CET 📍 In person at DFKI Kaiserslautern 🗣️ In English. Free. No AI safety background needed. 🎟️ Register: https://luma.com/nt0gvn82 #AISafetyKaiserslautern #MissionSafeKL #Kaiserslautern #AISafety
-
Technology is moving to the centre of global security discussions. Leaders from OpenAI, Anthropic and Hugging Face are due to brief the UN Security Council as governments consider the risks of increasingly powerful systems. A major shift in how the world approaches emerging technology. Read more in today’s Hive Signal.
-
Events over the past year – restricted access to frontier models, frontier model attacks, questions around the security of open source – mean that many governments now believe AI sovereignty is a zero-sum game. Cloudflare's answer: more local open-source models, model-agnostic security tools, and a commitment to giving nations genuine choice. Two new models are coming to Cloudflare: EuroLLM, which covers all 24 official EU languages, and Apertus, Switzerland's fully open model, trained on more than 1,500 languages. Both were built by public universities and research institutions. We're also launching hands-on workshops that help government cyber agencies and critical infrastructure operators build AI defenses that work with any model. The first runs in as part of Singapore International Cyber Week this month. In February, our CEO Matthew Prince told the India AI Impact Summit 2026 in New Delhi that decentralized, affordable access to AI is a matter of national resilience. The open source models from India, Japan and Singapore we added in 2025 were our first proof. Students, SMEs, civil society and the public sector used these models to build applications and services that impacted their local communities. Helping build a better Internet has always meant more options, not fewer. That's why we work on open standards that prevent vendor lock-in, why so much of what we build is free to start with, and why our network runs in more than 335 cities across 125+ countries, with GPUs for AI inference in more than 230 of them. We don't think any country should have to depend on one company for its AI. That includes us. https://lnkd.in/ew36wxgU Smrithi Ramesh Petra Arts Sebastian Hufnagel
-
The AI Scientist race is entering a new phase. Google’s new ScientistTwo is one of the most impressive steps toward autonomous scientific discovery I have seen recently. According to the published results, the system can identify research gaps, generate hypotheses, run experiments, and write papers that achieve review scores comparable to papers accepted at top AI conferences. It reportedly outperformed human state-of-the-art solutions on 86 out of 107 research problems (https://lnkd.in/dMkXFWpt). Unsurprisingly, an open-source Chinese reimplementation effort appeared almost immediately 😅, aiming to reproduce and extend the ScientistTwo approach (https://lnkd.in/dwZ58NsS) The next major AI battleground may not be chatbots or coding assistants, but autonomous scientific research itself. For universities, research institutes, and PhD programs, this is a trend worth following very closely over the coming months - especially for us. #AI #AIScientist #ScientistTwo #AgenticAI #ResearchAI #HigherEducation #FutureOfScience #Innovation
