From August 2, 2026, the European debate on synthetic content enters a new phase. Article 50 of the AI Act makes transparency obligations applicable to help people recognize when they interact with AI systems or encounter certain AI-generated or manipulated content.
This does not mean every retouched photograph needs a huge “AI” label. The rules distinguish between actors and situations: providers must make artificial origin machine-detectable in relevant cases, while deployers must inform people in situations such as deepfakes.
This is a cultural shift as well as a legal one. For years we discussed the quality of generative images: correct hands, believable movement, continuity and photorealism. Now another question moves to the foreground: what information should accompany an image when we can no longer infer its origin simply by looking at it?
This is where provenance systems such as Content Credentials enter. The C2PA standard can associate signed claims with an asset about its origin and transformations. It can record editing steps, tools and ingredients. It is not an absolute certificate of truth, but a verifiable history of the file.
For image and video makers, transparency can become part of the workflow. Camera, editing software, generative tools and platforms can contribute to the same content history. The credit would no longer be only a final line: it could be embedded in the technical life of the image.
Readability remains crucial. A perfect credential that nobody sees is of little use; a huge label applied without distinguishing full generation from a minor AI intervention can create confusion. Information design will therefore matter.
For journalism, cinema, advertising and digital art this does not have to limit experimentation. It can separate creative AI use more clearly from deceptive use. Declaring the process allows the work to be discussed for what it is without forcing the audience to play “real or fake?” all the time.
VISIBLE ORIGIN could become a new grammar for images. It will not automatically tell us what to believe, but it can restore something the generative era is making precious: the ability to know how an image arrived in front of our eyes.