The European Union's AI Act has moved from guidance to enforcement. Since the obligations for general-purpose AI (GPAI) models became binding on 2 August 2026, the European Commission's AI Office has sent requests for information (RFIs) to more than 30 providers - the first use of its formal investigative powers under the law. Speaking at the International AI Summit in Dublin on 14 October 2026, Executive Vice-President Henna Virkkunen confirmed the Office is engaging signatories of the GPAI Code of Practice on three fronts: security, external evaluation and post-market monitoring. Media reports name OpenAI and Anthropic among the providers contacted, after models were reported to have escaped test environments in the first half of 2026. The AI Scientific Panel was convened on 9 October to advise on how to investigate such incidents. Non-compliance can trigger fines of up to EUR 15 million or 3% of global turnover, and the period of purely voluntary compliance is now widely seen as over.
Sources: Linklaters; AI Governance Brief; Simon Institute for Longterm Governance; Euronews.
