The European Commission started enforcing new Artificial Intelligence Act transparency requirements on 2 August, requiring chatbots, AI agents, and avatars to tell users clearly that they are interacting with AI rather than a real person. The AI Office, working alongside national market surveillance authorities in each EU member state, took on enforcement duties for these obligations on the same date.
Under the rules, AI-generated or manipulated images, audio, and video that resemble a real person, object, place, or event (the European Commission's definition of a deepfake) must be clearly labeled and carry machine-readable marks so the content can be detected more easily. The same obligations cover emotion-recognition and biometric-categorization tools, plus AI-generated text published on matters of public interest without human editorial review.
National market surveillance authorities, the AI Office for systems under its direct supervision, and the European Data Protection Supervisor when an EU institution is itself a provider or deployer now share responsibility for policing compliance. Fines for noncompliant companies can reach €15 million or 3% of global annual turnover, whichever is higher; EU institutions, bodies, and agencies face a lower cap of €750,000, with proportionality applied for small and medium-sized companies, the European Commission said.
To help providers demonstrate compliance, the European Commission published guidelines and pointed to a voluntary Code of Practice on transparency of AI-generated content. More than 180 organizations have signed it so far, according to the European Commission. The AI Act entered into force on 1 August 2024, and its obligations continue to roll out in stages: these transparency rules are the latest tranche to take effect, with rules covering high-risk AI systems still ahead.
