European Commission adopts guidelines on AI Act transparency obligations

A client alert from the law firm Cooley explains that Article 50 imposes transparency duties on providers and deployers of certain AI systems of the EU AI Act (Regulation (EU) 2024/1689) from 2 August 2026. The European Commission adopted guidelines on these duties on 20 July 2026. Noncompliance can lead to fines of up to 15 million euros or 3 percent of worldwide annual turnover, whichever is greater. The Act reaches providers, deployers, importers and distributors worldwide when AI is placed on the EU market or its outputs are used there.

Article 50 covers four situations. Systems that interact directly with people, such as chatbots, voice assistants and AI agents, require providers to disclose the AI nature of the interaction unless obvious. Systems that generate or manipulate synthetic audio, image, video or text require providers to add machine-readable markings and a detection mechanism, with limited exceptions such as standard editing. Deployers of emotion recognition or biometric categorisation must inform the people affected. Deployers of deep fakes, and of AI-written text on public-interest topics, must disclose its artificial origin unless a person reviewed it substantively and took editorial responsibility.

The duties bind every in-scope system from 2 August 2026, whenever it reached the market, and earlier content need not be labelled retroactively. Generative AI systems already on the market have until 2 December 2026 to meet the marking and detection requirement. The AI Office has has also issued a voluntary Code of Practice for AI-generated content transparency, with icons for labelling. Signatories gain some presumption of conformity and gentler enforcement, while others face closer scrutiny. The firm advises businesses to inventory their systems, map them to these categories, and consider signing the Code.