Commission Tests AI Act Powers Over Escaped AI Agents
The Commission is testing whether the EU AI Act can reach AI agent incidents during development when they cause effects in the EU market.
AI Act
Commission Holds Special AI Scientific Panel Meeting
The Commission’s AI Scientific Panel will advise on frontier-model risks following recent loss-of-control incidents and increased EU AI Act scrutiny.
Commission Supports Slowing Frontier AI Development
Von der Leyen backed EU-led coordination to slow frontier AI development, using the AI Act’s systemic-risk rules and international security cooperation.
ENISA Tests Anthropic Mythos 5 and OpenAI GPT-6 Astra
ENISA’s testing of Anthropic Mythos 5 and OpenAI GPT-6 Astra offers an early test of AI Act supervision for systemic cyber risks.
Denmark Introduces AI Cheating Controls in Schools
Denmark will require oral defenses, supervised writing, and computer controls to curb AI cheating in upper secondary education.
Google Signs EU AI Content Transparency Code
Google’s voluntary AI-content-labeling commitment may support AI Act compliance, but providers remain responsible for meeting Article 50 transparency duties from August 2026.
Commission Launches EU AI Cybersecurity Action Plan
The Commission’s AI cybersecurity Action Plan coordinates EU actors to manage risks from advanced AI models and strengthen digital resilience.
Leading AI Models Show EU Compliance Gaps
Aithos found leading AI models often breach GDPR and the EU AI Act in agent tasks, exposing deployers to fines and requiring real-world compliance testing.
EU Seeks Access to Anthropic Mythos Over Cybersecurity Risks
The EU is moving closer to access to Anthropic’s Mythos AI model amid cybersecurity, AI governance, and tech sovereignty concerns.
Commission Issues Draft Guidance on High-Risk AI Classification
The Commission’s draft AI Act guidelines explain how to identify high-risk AI systems, provide practical examples, and confirm new enforcement timelines following the AI Omnibus agreement.