EU
European Commission / AI Office
Primary enforcer of the transparency rules and GPAI provisions; can now inspect general-purpose AI models, restrict EU market access, and levy fines. The AI Office has 145 staff, with 34 on regulation/compliance and 38 on safety teams.
General-purpose AI model provider now under direct EU enforcement scrutiny; signed the voluntary GPAI Code of Practice; offered the EU access to its 'GPT-5.5-Cyber' model amid cyberattack concerns.
General-purpose AI model provider under new enforcement powers; signed the GPAI Code of Practice; agreed to give EU cybersecurity agency ENISA access to its 'Mythos' model after Commission officials visited San Francisco.
General-purpose AI model provider now under enforcement scrutiny; signed the EU Code of Conduct on transparency; warned that overlapping AI labels could make it harder for people to get clear context.
Declined to sign the safety chapter of the GPAI Code of Practice; already deploys 'AI Info' labels on Instagram and Facebook.
Already requires AI-generated content labels ahead of enforcement; has labeled over 3 billion items.