🛡 EU begins enforcement of EU AI Act: first requests to OpenAI, Anthropic and Google
The European Commission's AI Office sent information requests (RFI) to general-purpose model providers on August 29, 2026 — according to Euractiv, OpenAI, Anthropic and Google received them. They are asking about model protection from attacks, external evaluation, and post-release monitoring.
🌍 Refusal or inaccurate response — fine up to €15 million or 3% of global turnover, and responses will supplement a permanent supervisory file. GPAI provider obligations have been enforceable since August 2 — the first requests were sent within four weeks.
👤 Nothing has been announced about the unavailability of models in the EU — these are forecasts, not decisions. The requests are addressed to companies, not individuals: a locally launched model is outside jurisdiction.
Source 1: https://tokenstead.ai/guides/eu-ai-act-first-enforcement-security-rfis Source 2: https://www.cnbc.com/2026/08/03/eu-ai-act-enforcement-powers.html