ChatGPTCast journal
What the EU AI Act transparency rules mean for the media you see
The European Union's AI Act includes a set of transparency duties in its Article 50. They have applied since 2 August 2026, and they affect what people in the EU should be told when they talk to an AI system or watch synthetic media. They are worth understanding even outside Europe, because many services apply the same labels everywhere.
By ChatGPTCast editorial · · 2 min read
Know the four main duties
First, chatbots and similar systems must make it clear that people are dealing with AI, unless that is already obvious to a reasonably well-informed person. Second, providers of generative AI must mark synthetic text, images, audio and video in a machine-readable way so it can be detected. Third, those deploying emotion recognition or biometric categorisation must inform the people exposed to it. Fourth, deployers must disclose deepfakes, and must disclose AI-generated text published to inform the public on matters of public interest.
For providers of generative systems already on the market before 2 August 2026, a transition period for machine-readable marking runs until 2 December 2026.
Notice the exceptions
The rules carry exceptions. Purely personal, non-professional use falls outside them, and so do systems developed only for scientific research. Content only lightly edited with AI assistance is treated differently from content that is substantially altered. AI-generated public-interest text does not need the disclosure when it has gone through human review or editorial control and a person or organisation takes editorial responsibility for it.
The European Commission published final guidelines on these obligations in July 2026, and a voluntary code of practice covers marking and labelling. Expect the practical details, such as icons and wording, to keep developing.
Use the rules as a viewer
A machine-readable mark is meant for software, so you may not see it. What you can do is notice visible disclosures, read them, and ask a service how it marks AI content if it does not say. For anyone planning AI demonstrations, the safest habit is to label generated media plainly and to say when a viewer is talking to an automated system.
- Expect chatbots in the EU to say they are AI.
- Look for deepfake disclosures on realistic synthetic media.
- Remember that machine-readable marks are often invisible.
- Check exceptions before assuming a label is missing unlawfully.
- Label generated media clearly in your own demos.