Overview
- On 2 August 2026 the AI Act’s transparency and labelling provisions came into force, obliging providers to mark many AI‑generated or AI‑manipulated images, text, audio and video and to inform users when they interact with chatbots.
- The law requires visible labels plus machine‑readable metadata or watermarks and forces disclosure of automated emotion or biometric recognition, while leaving the EU Commission power to investigate, request documents, fine providers and, in extreme cases, block models from the EU market.
- Germany has named the Bundesnetzagentur as its national enforcement and coordination authority to handle market surveillance, complaints and guidance for businesses and consumers.
- Major gaps remain: the Commission’s proposed icons are non‑binding, many editorial and private‑user exceptions apply, and postponed rules for high‑risk systems and older general‑purpose models push substantive obligations to 2 August 2027.
- Experts warn labels only reveal whether AI was used and cannot certify accuracy or fairness, they can be gamed or ignored, and inconsistent platform implementations risk fragmenting user understanding and limiting the rules’ effectiveness.