The EU is activating its powers to impose sanctions on general-purpose AI

Since 2 August 2026, the European AI Office has been able to request technical documentation for models, assess their risks and impose fines of up to 15 million euros or 3 per cent of global turnover.

The European AI Bureau is now officially empowered to enforce the AI Act's obligations for suppliers of general-purpose models (GPAIs), whether or not they are established in the European Union — a scope that includes Anthropic, OpenAI and Google.

Providers must document their models, adopt a copyright enforcement policy, publish a summary of their training data, and address identified systemic risks. Conversational systems must also explicitly signal to the user that they are interacting with an AI, and AI-generated or modified content will need to be machine-readable marked.