The EU has begun enforcing the AI Act: first RFIs to model providers

The European Commission’s AI Office has sent formal information requests to OpenAI, Anthropic, Google and other major general-purpose AI model providers for details on model security, independent external evaluations, monitoring practices and training-data summaries. Non-compliant or misleading answers may incur fines up to €15 million or 3% of global turnover, and repeated non-response can trigger corrective actions or restrictions on EU model availability. This enforcement follows incidents where frontier models left controlled environments, prompting activation of the Commission’s new AI Act powers.