New EU rules requiring some artificial intelligence systems to clearly identify AI-generated or manipulated content came into effect on 2 August 2026.
The measures respond to the growing difficulty of distinguishing AI-generated and altered material from authentic, human-made content, with risks including misinformation, fraud, impersonation and consumer deception, the European Commission said in a statement on Sunday.
Under the new transparency obligations, certain AI-generated or manipulated images, audio and video that resemble real people, objects, places, entities or events — often referred to as “deepfakes” — must be clearly and visibly labelled and include machine-readable marks.
The rules also cover emotion recognition and biometric categorisation tools.
Text published to inform the public on matters of public interest must also be labelled where there has been no human review or editorial control.
Users must also be clearly told when they are interacting with an AI system rather than a real person — such as a chatbot, AI agent or avatar — under the transparency rules.
Who enforces the rules and what are the penalties?
Enforcement will be carried out by national market surveillance authorities, the Commission’s AI Office for systems under its supervision, and the European Data Protection Supervisor when EU institutions are providers or deployers.
Fines for companies can reach up to €15 million or 3% of global annual turnover, while EU institutions, bodies and agencies can be fined up to €750,000.
The Commission said proportionality would be taken into account for small and medium-sized enterprises and small mid-cap companies.
The transparency obligations are part of the EU’s AI Act, which entered into force on 1 August 2024 and is being applied in stages.
