New disclosure mandates force chatbots, deepfakes and generative tools to reveal their artificial origins across the bloc.
From today, 2 August 2026, the European Commission’s AI Office and national authorities begin enforcing key parts of the Artificial Intelligence Act. Transparency obligations under Article 50 now apply in full.
Chatbots and interactive systems must tell users they are dealing with AI, not a human. Providers of generative systems that create or alter text, images, audio or video must mark outputs in machine-readable form so detection tools can spot them. Deployers must label deepfakes and disclose AI-generated text published to inform the public on matters of public interest. Emotion recognition and biometric categorisation systems also require clear notice to the people exposed to them.
What the rules require in practice
The obligations cover four situations and apply to almost any AI system used in them, not only high-risk ones. Notification must come at the first interaction or exposure and remain clear and distinguishable. Vague footer text or fleeting labels will not suffice.
A Code of Practice on AI-generated content, signed by more than 180 organisations, offers practical guidance on marking and labelling. The Commission has also issued guidelines to help providers and deployers interpret the rules. Systems already on the market before today receive a short extra window until 2 December 2026 for the machine-readable marking requirement.
Enforcement and penalties
The AI Office oversees general-purpose AI models and certain integrated systems. National authorities handle most other cases. The European Data Protection Supervisor covers EU institutions. Fines for transparency breaches can reach €15 million or 3 percent of global annual turnover, whichever is higher. Proportionality applies for smaller firms.
Prohibited practices already in force continue under active supervision. High-risk system obligations were delayed by the AI Omnibus to December 2027, giving companies more time on the stricter conformity rules.
Why this matters now
AI tools have moved deep into daily life. The new rules aim to cut deception, let people make informed choices, and give businesses clearer compliance paths. Users should start seeing more explicit AI labels on chat interfaces, synthetic media and public-interest content in the coming weeks. Companies that fail to adapt risk both financial penalties and loss of trust.