You just finished rendering a gorgeous digital model for your next social media ad campaign. The lighting is perfect, the skin texture looks completely real, and best of all, you didn't have to pay for a casting call, studio time, or talent residuals.
Then your legal team drops a bomb on your desk. If that ad runs in New York, you have to explicitly label that fake person.
New York General Business Law Section 396-b takes effect today, June 9, 2026. It marks a massive shift in how brands use artificial intelligence. From this point forward, any commercial advertisement distributed in the state that features a computer-generated human must feature a clear disclosure. If you try to pass off a fake person as a real human actor, you face immediate civil penalties.
This isn't a vague guideline or a suggestion for best practices. It's a strict statutory mandate. If you're a brand manager, an agency director, or an independent creator, you need to understand exactly what this law requires before you launch your next campaign.
What Counts As A Synthetic Performer
The state of New York defines a synthetic performer as a digitally created asset that is created, reproduced, or modified by a computer using generative AI or a software algorithm. The core test relies on intent. If the asset creates the impression that an actual human is engaging in a visual or audiovisual performance, it falls under the law.
This definition covers a lot more ground than most creative departments realize. We aren't just talking about highly advanced deepfakes of real people. The law specifically targets digital assets that are not recognizable as an identifiable natural person.
- Full-body AI models: Those photorealistic AI fashion models you see modeling clothes on e-commerce sites.
- Digital extras: Computer-generated background actors used to fill out a crowd scene in a commercial video.
- Partial human features: A digitally generated hand modeling a watch, or an AI-generated set of teeth in a toothpaste ad.
The phrase "software algorithm" is causing serious panic in agency legal departments. The statute doesn't limit itself to tools like Midjourney, Runway, or Sora. If you use traditional, non-AI visual effects software to render a human-like figure from scratch, you've still created a synthetic performer. If a consumer looks at the ad and assumes they're looking at a real human being, you have to label it.
The Costs Of Hiding Your AI Talent
The financial penalties under Section 396-b stack up fast. A first-time violation hits you with a $1,000 civil penalty. Every subsequent violation carries a $5,000 fine.
Think about how modern ad operations function. You don't just run one ad; you run dozens of variations across different platforms. If you run a multi-variant programmatic campaign featuring an unlabeled synthetic model across ten different ad sets, those "subsequent violations" can wipe out your profit margins in a weekend.
The legal burden falls squarely on the entity that creates or commissions the ad. The law includes an actual knowledge qualifier. If you know you used a synthetic performer, you must disclose it.
Publishers, media outlets, and platforms get a free pass here. Traditional television networks, streaming services, social media platforms, and billboard companies aren't liable if they accidentally run an unlabeled ad. New York isn't going after the companies hosting the media. They're going straight after the brands and creative agencies.
The Safe Harbors Where Disclosures Aren't Required
The law isn't a blanket ban on AI, and lawmakers did carve out a few notable exceptions to prevent a total freeze-out of the creative industry. You don't need a label in these specific scenarios:
Expressive Works and Entertainment Promo
If you're creating promotional materials for a movie, a television show, a streaming series, or a video game that uses synthetic performers within the work itself, you're exempt. An ad for a sci-fi film featuring digital actors doesn't need an ugly disclaimer stamped on the trailer, provided the ad accurately reflects the content of the expressive work.
Audio-Only Advertisements
The current iteration of the law specifies visual and audiovisual performances. Radio ads, podcast sponsorships, and digital audio spots that use AI-generated voices do not trigger the disclosure requirement under this specific statute. Be careful, though; other state right-of-publicity laws still protect real people from unauthorized voice cloning.
Basic Language Translation
Using AI to translate the voice or mouth movements of a real human actor into another language is completely fine. If a real person filmed the ad and you used algorithmic tools to dub their voice into Spanish, you don't need a synthetic performer disclosure.
How To Draft A Compliant Disclosure
The statute demands a "conspicuous disclosure," but it fails to provide an exact script or font size. It leaves the specific mechanics open to interpretation, which means early enforcement will likely lean on established Federal Trade Commission standards.
To keep your brand safe, your disclosures should be impossible for a consumer to miss. Don't hide the text in a massive block of terms and conditions at the bottom of a web page. Don't use a tiny grey font against a white background.
Many brands are choosing to use explicit terms like "AI-generated image" or "Digital performer used" directly on the creative asset. For video ads, this means placing clear, high-contrast overlay text during the initial seconds of the clip. For static social media ads, the label needs to live directly within the frame of the graphic.
Your Action Plan For Creative Compliance
If you want to avoid a compliance nightmare, you need to update your operational workflow immediately. Relying on your creative team to voluntarily flag every AI asset isn't a viable long-term strategy.
First, audit your active asset library. Identify every creative asset currently running in the Northeast region. If an ad features a human face, body part, or silhouette generated by a computer, pull it down or add the disclosure text.
Second, rewrite your agency vendor agreements. If you outsource your creative production to third-party agencies or independent contractors, your contracts must include strict representations and warranties regarding generative AI. Force your partners to explicitly disclose if they use synthetic performers, and build indemnification clauses into your contracts so your brand isn't left holding the bag for an agency's secret AI usage.
Finally, integrate a mandatory compliance checkpoint into your creative sign-off process. Before any ad goes live, the media buying team must verify the origin of the assets. Treat the synthetic performer check with the same gravity you treat trademark searches and copyright clearances. The age of frictionless, unlabelled digital humans is officially over.
,webp/027/070/476/1280x720.17542290.jpg)
