Since 2 August 2026, some transparency obligations under the EU AI Act have applied across the European Union. They cover specific situations, such as informing people when they are talking to an AI system and identifying certain manipulated content that appears authentic. They also cover some AI-generated texts on matters of public interest, with exceptions where there is human review or editorial control and responsibility for the publication.[1][2]
I find it helpful that transparency takes this concrete form, in situations that can affect what we see and how we interpret it. If an image looks real or an answer sounds confident, knowing AI was involved can help us put it in context.
A label still leaves me with questions. Is the content accurate? Who is responsible if something is wrong? What criteria did the system follow? The notice might also go unnoticed if it is hidden or confusingly worded.
Without knowing that AI was involved, we are missing something we need to assess what is in front of us. I wonder what we expect companies to tell us when they use these tools in their communication. Is a notice enough? Or do we also need to know what a person has checked?
Eugénie Morant Vermeersch