What happened
From 2 August 2026, the European Commission's AI Office, together with national market surveillance authorities, begins formally enforcing the AI Act's Article 50 transparency obligations — covering AI systems that interact with people, generate synthetic content, perform emotion recognition/biometric categorisation, and deepfakes. This is distinct from the earlier-reported 'Digital Omnibus' (Regulation 2026/1744), which postponed high-risk Annex III rules but explicitly did NOT defer Article 50. The Commission published a first list of 180+ organisations (including Amazon, Google, Microsoft, Anthropic, OpenAI, Mistral) that signed the voluntary Code of Practice on transparency of AI-generated content ahead of the enforcement date, and national competent authorities begin supervision/enforcement powers on 2 August 2026.
Why it matters
This is the first live enforcement window for substantive AI Act obligations beyond prohibited practices and GPAI rules — penalties reach up to €15 million or 3% of global annual turnover, enforced by national market surveillance authorities. Any provider or deployer of generative AI, chatbots, or synthetic-media tools serving EU users must now have marking/labelling and disclosure mechanisms operational; multiple analysts (Reed Smith, Forkast) flag that agent-specific disclosure (Art. 50(1)) is NOT covered by the Code of Practice, leaving a compliance gap for the fast-growing AI-agent market.
Action needed
Providers/deployers of GPAI, generative AI, or synthetic content systems reaching EU users must implement Article 50 disclosure and machine-readable marking now; non-signatories to the Code of Practice must demonstrate compliance through other means, assessed by national market surveillance authorities.