Skip to content
Start a project

AI content labelling rules are converging

Regulators, ad codes and platforms are moving towards disclosing AI-generated content, with C2PA Content Credentials as a common thread. What is settled, what isn't, and what to ask your studio.

By Brand1 studio7 min read

A roadside billboard printed with clouds, set against a real blue sky
Header photograph, used for mood.Photo: John Maldonado / Unsplash

Key takeaways

  • Lawmakers, ad standards bodies and platforms increasingly expect realistic AI-generated images, video and audio to be disclosed.
  • Exact label wording, size and placement still vary by market and platform, so expect guidance to keep changing.
  • C2PA Content Credentials add a tamper-evident record of how a file was made; they support a visible label but do not replace it.

If you approve advertising, the question is no longer whether AI-made work needs a label. It is which label, where, and who keeps the record. The answers differ by market, but they are moving in the same direction.

Do AI-generated ads need a label?

Increasingly, yes. Lawmakers, advertising standards bodies and the major platforms now expect realistic synthetic images, video and audio to be disclosed to the people who see them. The European Union's AI Act sets transparency duties for AI-generated and manipulated content, including deepfakes. The large social platforms ask advertisers to declare realistic AI content and apply labels of their own.

Disclose what is synthetic. Keep the record. Be able to prove both.
9:41@onestory.filmOne story60sMASTER60 sCUTDOWNS30 · 15 · 6VERSIONS32Onestory.@onestory.film60 sOnestory.@onestory.filmStoryboardShot list · v3 · 6 of 2401 · Dark02 · First light03 · StreakOne story.04 · HorizonOne story.05 · TitleOne story.06 · End card60 sOnestory.@onestory.film60sMASTER60 sCUTDOWNS30 · 15 · 6VERSIONS3260 s
Launch film social cut-downs, specimen project

What is still unsettled?

The exact wording, size and placement of a label still vary. So does the line between a light retouch and synthetic content. Expect guidance to keep changing for a while, and expect regulators to judge intent and effect, not only the presence of a badge.

What are Content Credentials (C2PA)?

Content Credentials, built on the open C2PA standard, attach a tamper-evident record of how a file was made. Several camera makers, creative tools and platforms already read or write them. They do not replace a visible label, but they make it far easier to show your working.

What should you ask your studio?

Which tools made each asset? Is a visible label applied where the market requires one? Do the delivered files carry Content Credentials? Who keeps the production record, and for how long? A good studio answers all four without being asked. This piece is a guide, not legal advice.

Our AI and Content Credentials policyAll Journal pieces
FAQ

Questions on this piece

Do AI-generated ads need a label?

In a growing number of markets and on most major platforms, realistic AI-generated or AI-altered images, video and audio in advertising are expected to be disclosed. Rules differ on wording and placement, and light retouching is often treated differently from synthetic content. Brand1 labels AI-made work where a market or platform requires it and keeps a production record for every asset. This is guidance, not legal advice.

What are Content Credentials?

Content Credentials are tamper-evident metadata, built on the open C2PA standard, that record how a digital file was made and edited, including whether AI tools were used. Cameras, creative software and platforms can write and read them. They complement a visible AI label rather than replace it. Every file Brand1 delivers carries signed Content Credentials listing the AI tools used and the human edits.

What does the EU AI Act say about AI-generated content?

The EU AI Act sets transparency duties for AI-generated and manipulated content. Providers of generative AI systems must mark outputs in a machine-readable way, and people who publish deepfakes must disclose that the content is artificial, with lighter rules for clearly artistic or satirical work. Detailed guidance and codes of practice continue to develop, so check current advice for each campaign.

Next in the Journal

6 min read
1brand150°CASE STUDY · 01Brand first.Then everyexpression.23
Brand guidelines for AI: why the PDF stopped workingA PDF can't brief an AI tool. What it takes to turn a brand book into something every person and every model actually uses.