The European Union’s AI Act has officially entered its enforcement phase as of April 1, 2026, with the newly established European AI Office beginning compliance audits of companies deploying high-risk and general-purpose AI systems within the bloc. The first round of audits will focus on frontier model providers — including OpenAI, Google DeepMind, Anthropic, and Meta — with preliminary findings expected by June and potential fines by late summer.
The timing is particularly significant given the Claude Mythos leak, which exposed internal safety documents that regulators are now reportedly reviewing. Under the AI Act, companies deploying general-purpose AI models with systemic risk face fines of up to 3% of global annual revenue for non-compliance. Several AI companies have already established EU-based safety teams and published transparency reports in preparation, but industry lawyers warn that the Act’s requirements around “adequate safeguards” remain vaguely defined, setting the stage for legal challenges.