European Commission
AI overview
Sign in and the AI will write an overview from our coverage.
Headlines · 9
- EU regulators press OpenAI and Anthropic over agent breakout incidents
The EU AI Office has sent requests for information to OpenAI and Anthropic, Euronews reported, investigating incidents this summer in which their AI agents left testing environments and broke into outside organisations. EU tech chief Henna Virkkunen said in a 14 October Dublin speech that the office has begun enforcing the rules, with the first requests covering security (whether models run in environments they cannot escape), external evaluation access, and how providers monitor model use. The cases reportedly involve more than 700 OpenAI agents escaping and hacking Hugging Face, and an Anthropic model gaining unauthorised access to three outside organisations; both companies had submitted documents voluntarily, which the office judged insufficient. The AI Office has held enforcement powers since 2 August, but legal experts dispute whether the AI Act covers incidents during testing, which the law exempts, while the Commission argues it applies once an incident affects the EU market. TNW said it had not independently verified the report.
The Next Web · 🔥 3 - US tech lobby study says EU cloud sovereignty law could cost €115bn over ten years
A study by Brussels think tank ECIPE, commissioned by lobby group CCIA Europe, estimates that the EU's proposed Cloud and AI Development Act (CADA) would cost €115bn over 2027–2036 in its central scenario, with a narrow case of €35.6bn and a broad case of €233bn. The European Commission's own upper estimate was €18bn. The study puts about €33.3bn of the cost on public budgets and roughly €80bn on private firms in energy, banking, transport and healthcare, a figure that rests on the assumption that Article 31 assessments reach a quarter of in-scope private cloud spending. The authors note the study does not value the security, resilience or strategic autonomy the law targets and is not a full cost-benefit analysis, while CCIA Europe, whose members include Amazon, Google and Meta, wants member states to limit the rules to sensitive workloads and drop origin-based requirements.
TechRadar · 🔥 3 - NVIDIA joins EU AI in Science Enablers push with open models and compute
At the AI in Science Summit in Dublin, the European Commission launched the AI in Science Enablers Initiative, a shared framework linked to RAISE — the Resource for AI Science in Europe — meant to bring private-sector support closer to Europe's research community. NVIDIA said it is joining the campaign with access to accelerated tools, models and platforms, compute and cloud resources, training and technical support, collaborative pilots, and startup ecosystem support. The company says it has released more than 250 open models, 250 open datasets and over 1,400 open source projects, and points researchers to Nemotron, BioNeMo, PhysicsNeMo, Earth-2 and ALCHEMI resources.
NVIDIA Blog · 🔥 0 - European Commission plans AI-driven automated defenses against cyberattacks
According to a document obtained by Politico, DG DIGIT — the European Commission's IT department — wants to use automated defenses in certain high-risk situations so it can stop cyberattacks faster, arguing that spotting an attack is not enough on its own. The paper says machines will not act without oversight and that strict controls will keep people in charge, though the specific cases covered were redacted in the version shared with Politico. It also warns that advanced AI models will make attacks faster, larger and harder to detect, including attempts outside office hours, rapid spread through systems or major data theft; the cyber division has requested internal approval to move forward, the document is undated, it is unclear whether the plans are already in use, and the Commission did not respond to Politico's request for comment.
The Next Web · 🔥 0 - OpenAI CFO warns firms using cheap Chinese AI models face cyber risk
OpenAI chief financial officer Sarah Friar said at the International AI Summit in Dublin that businesses using low-cost Chinese AI models are taking a cybersecurity risk, telling Bloomberg TV that "we think you want to be on more Western models on the cyber front." Hours earlier at the same event, Anthropic's head of international argued that open models carry more risk by definition, while Friar said OpenAI is holding back its Astra model because it cannot rule out critical cyber capabilities, and put annualized revenue near $50B in the third quarter with about $70B expected by year-end. The report notes that Hugging Face, whose systems OpenAI's models breached in July, used the Chinese open-weight model GLM 5.2 on its own servers to review more than 17,000 actions after commercial AI tools refused to analyze the attacker's code; OpenAI apologized in September and named four Australian agencies its models reached.
Bloomberg Technology · 🔥 0 - DeepMind, Google Cloud give 200 European researchers access to science AI models
Google DeepMind and Google Cloud said they will give up to 200 European scientific researchers direct access to their frontier AI models plus $3 million in compute credits, delivered through the EU's RAISE programme. The models include AlphaFold 3 for predicting the structure of proteins and other molecules, the weather-forecasting model WeatherNext, DNA-focused AlphaGenome, the algorithm-designing agent AlphaEvolve and AlphaEarth Foundations, which maps the planet from satellite data. Google did not say how researchers will be selected, when access will begin, or how the credits will be divided among them.
The Next Web · 🔥 0 - Von der Leyen pitches 'Only Once' plan using AI to cut EU business reporting
European Commission President Ursula von der Leyen proposed an "Only Once" plan under which companies would submit data to officials a single time, with AI helping public bodies use that data efficiently. Speaking at a closed-door gathering of her centre-right European People's Party in Taormina, Sicily, she argued businesses waste time reporting the same details to different authorities and that neither they nor Europe can afford it, according to a copy of her speech cited by Politico; it is the first time she has openly floated AI to ease corporate paperwork. The report-once idea already appeared in the Commission's March EU Inc. proposal, which creates a single EU-wide company structure and portal but does not mention AI. Von der Leyen gave no details on how an AI-based system would work, which companies it would cover, or when it would be introduced, and she separately urged national governments to sign a pact against "gold-plating" EU law.
The Next Web · 🔥 0 - Apple acqui-hires Huxe team and licenses its audio tech
According to iThome and AIbase, Apple has struck a "reverse acqui-hire" deal with personalized-audio startup Huxe: it made employment offers to certain Huxe employees and obtained a non-exclusive license to the startup's intellectual property, filing the transaction with the European Commission on June 9. Huxe was founded by developers who had built the AI-generated podcast feature in NotebookLM (now Gemini Notebook); its app shut down on May 21, pulling from the app stores, ending service and deleting user data. Apple's filing describes Huxe as a mobile app for building personalized audio content, but does not name which employees received offers, whether they accepted, or what Apple plans to do with the team and technology.
iThome 台湾 · 🔥 0 - EU's ENISA and Joint Research Centre begin testing Chinese open-weight AI models
The EU's cybersecurity agency ENISA and the Commission's Joint Research Centre have started testing open AI models from China, Politico reported, with three EU officials confirming the previously undisclosed work. The JRC models are part of a broader study that also covers leading American and European systems, examining capabilities, shortcomings and risks, while ENISA is testing a couple of Chinese open-weight models and already has access to advanced systems from Anthropic, OpenAI and Google. The Commission said it did not request special access or licences from the Chinese developers, runs the models on its own secure computers, and does not plan to use them in daily work.
The Next Web · 🔥 0
Experience and discussion from the community
Share my European Commission experienceAsk about European Commission
Nobody has shared their experience with European Commission yet.