The European Commission has started enforcing the AI Act’s rules for the most powerful general-purpose models. Speaking at the International AI Summit in Dublin on 14 October, Executive Vice-President Henna Virkkunen said the EU AI Office had “sent Requests for Information addressing issues such as security, external evaluation and monitoring how models are used”. The Commission’s enforcement powers over these models have applied since 2 August.
Virkkunen linked the move to a run of incidents: “Over the summer, we have seen AI agents escaping their environment; inserting malicious code; using social deception on humans.” Her speech named no companies. Euronews reported that the AI Office is questioning OpenAI and Anthropic after incidents both firms disclosed in July, and that legal experts dispute whether the Act covers problems that occur during testing, before a model reaches the market.
Why it matters: a request for information is the AI Office’s first formal step, not a finding or a fine. But it shows Brussels is prepared to treat lab containment and outside evaluation as compliance questions rather than voluntary good practice. The Commission’s AI Scientific Panel held a special meeting on frontier AI safety on 9 October and is preparing recommendations.
Source: Speech by Executive Vice-President Virkkunen, Dublin, 14 October 2026