New York’s synthetic performer disclosure law was signed in December 2025 and took effect on 9 June 2026. It requires anyone who produces or creates an advertisement to identify when that ad includes an AI-generated synthetic performer. The state describes it as the first law of its kind in the country. Penalties run to $1,000 for a first violation and $5,000 for each one after.

What actually counts as a synthetic performer

This is where most coverage is vague, and where the practical risk sits.

A synthetic performer is a digitally created asset, produced or altered by generative AI or comparable software, intended to create the impression of a human performance by someone who is not recognisable as any identifiable real performer. The operative idea is a fabricated human being who never existed.

That definition sorts things more cleanly than most teams expect:

• An AI-generated presenter who looks like a person but is nobody in particular: covered.
• A fabricated human voice reading your script in a video ad: covered, since the law reaches visual and audiovisual advertising.
• A digital replica of a real, identifiable person: a different problem, governed by separate rules on likeness and digital replicas.
• AI-generated product shots, backgrounds, or an animated logo with no human performance: not a synthetic performer.
• A real actor whose footage you colour-graded or upscaled with AI tools: not a synthetic performer.

The dividing line is human performance, not AI involvement. A heavily AI-assisted ad containing no fake human does not trigger the rule. One AI spokesperson in an otherwise conventional ad does.

Why “we are not in New York” is not an answer

The obligation attaches to advertisements distributed to a New York audience, whatever the advertiser’s location. If you run Meta or YouTube video without geographic exclusions, you are distributing to a New York audience. A four-person agency in Manchester producing video for a US client is inside the scope the moment that ad serves in the state.

Worth noting, because it explains why so few marketers have registered this: almost all the compliance writing on the law is aimed at film and television production companies. That is the industry it was framed around, and the industry SAG-AFTRA campaigned for. The wording is broader than the framing.

The process, which takes about an hour to set up

Consider a five-person ecommerce brand that started using an AI presenter in TikTok and YouTube ads in March because paying a creator for every product launch was not affordable. Nothing about that decision was wrong. The gap is process, not judgement.

1. Add one question to creative sign-off: does this ad contain a human face or voice that no real person performed? If yes, it needs a disclosure.
2. Write one standard disclosure line and reuse it everywhere. Conspicuous means a viewer can actually notice it, so two frames of grey text at the end will not do the job.
3. Keep a simple record of which assets are synthetic. This is dull, and it is the thing that protects you if anyone ever asks.
4. Go back through your ad library. Creative produced before June is still running.

The realistic failure mode is not a fine. It is that nobody at a small company owns the question, so it falls between the freelancer who produced the ad and the founder who approved it, and neither one knows the question exists.

When this does not apply to you

If your video advertising is a founder talking to a phone camera, there is nothing to do. If you use AI only for editing, captions, thumbnails, or product photography, there is nothing to do. Adding a disclosure where no synthetic performer exists is not a harmless precaution either: it confuses viewers and, in a category where trust is part of the sale, tells them something untrue about your work.

This also depends on how you advertise. If everything you run is static image or text, the rule as written is about visual and audiovisual performance, and you are outside it.

The pattern worth noticing

New York went first, and the interesting detail is what it chose to regulate. Not AI use. Not AI-generated content generally. Specifically the fabricated human. Regulators are converging on the view that the harm is impersonating humanity rather than automating work, which is the same instinct behind the transparency provisions in the EU AI Act disclosure rules. If you are building an internal process, build it around that question rather than around a list of approved tools. The tool list will be obsolete by Christmas. The question will not.

Practical next step: open your ad account, filter video creative made since March, and count how many pieces feature a person who does not exist. If the answer is above zero, that is this week’s job. The state’s own summary of the law is in the New York announcement on synthetic performer disclosure.

Editor’s note: This area changes quickly, so check the latest platform policy before making compliance decisions.

The post New York Has Required AI Performer Disclosure Since June. Most Advertisers Still Have No Process. appeared first on Mark8ng.com.