How can you tell that a text, image or voice was produced by artificial intelligence, or that the party you are dealing with on an online service is a machine? To preserve trust, the AI Act (Regulation (EU) 2024/1689) introduces transparency obligations, mainly in its Article 50.
Knowing you are talking to an AI
When an AI system is designed to interact directly with people — a chatbot, for example — users should in principle be informed that they are communicating with a machine, unless this is obvious from the context.
Identifying AI-generated content
Providers of generative AI must ensure that artificially generated or manipulated content is identifiable, in a machine-readable format. In addition, certain content must be clearly disclosed: this is the case for deep fakes — images, audio or video resembling real people or events — as well as text published to inform the public on matters of public interest.
Who bears these obligations?
These obligations fall on both providers (who design the systems) and deployers (who use them), for example whoever disseminates a deep fake. The aim is to help everyone know when content or an interaction is artificial, so that people can weigh it accordingly and keep an informed sense of trust.
The case of generative AI
According to the European Parliament, generative AI (such as large language models) is not classified as high-risk, but must comply with transparency requirements and EU copyright law: disclosing that content was generated by AI, designing the model to prevent it from producing illegal content, and publishing summaries of the copyrighted data used for training.
A gradual entry into application
According to the European Commission, the AI Act's transparency rules take effect in August 2026. To clarify how they apply, the Commission presented a voluntary Code of Practice on the marking and labelling of AI-generated content (June 2026) and is preparing guidelines on transparent AI systems, expected in the second quarter of 2026.
The technical modalities (marking, watermarking, metadata) are still being clarified through guidelines and codes. It is therefore prudent to follow official publications and to date any information, as the framework is evolving.