On June 9, 2026, New York became the first U.S. state to require advertisers to tell viewers when an ad features a fabricated human. The law — S.8420-A/A.8887-B, signed by Governor Kathy Hochul in December 2025 and codified as an amendment to New York General Business Law § 396-b — directs any advertiser who, with "actual knowledge," includes a "synthetic performer" in a commercial ad to "conspicuously disclose" that fact. Penalties run to $1,000 for a first violation and $5,000 for each one after. The statute defines a synthetic performer as a digitally created asset "created, reproduced, or modified by computer, using generative artificial intelligence or a software algorithm," made to look like a human performer who is not any identifiable real person.
This is the influencer-economy question arriving in statute. Brands increasingly deploy AI-generated spokespeople and virtual creators that look, talk, and post like humans but answer to no agent and never age out of a campaign. New York's premise is that a viewer deserves to know whether the face selling them a product is a person or a render.
The case for disclosure is strong
The strongest argument for the law is the simplest: it is a transparency rule, not a ban. It does not prohibit synthetic performers, restrict who may make them, or dictate creative choices. It asks only for a label. Disclosure-based regimes are the least speech-restrictive tool a regulator has, and the federal baseline already points the same direction. The Federal Trade Commission's revised Endorsement Guides (16 CFR Part 255), finalized in 2023, treat a fabricated or computer-generated endorser as an endorsement subject to the same truthfulness and material-connection rules as a human one — a fake persona vouching for a product is still a representation to consumers. New York's law is continuous with that logic. SAG-AFTRA, which backed the bill, frames it as protection for both consumers and a creative workforce being asked to compete against synthetic stand-ins. Governor Hochul cast it as keeping the state "at the forefront of responsible innovation." On principle, none of that is objectionable. Consumers benefit from knowing what they are looking at, and a clear, narrow disclosure duty is exactly the kind of proportionate intervention a pro-innovation regulator should prefer over content bans.
The carve-outs show genuine drafting care. The law exempts audio-only ads, ads where AI is used solely to translate a real performer's speech, and promotional material for expressive works — films, games, streaming — where synthetic characters are part of the work itself. Those exemptions track real distinctions and blunt the most obvious overreach objections.
Where the design breaks down
The problem is not the goal but the mechanics. Three features make the law harder to comply with than it should be.
First, the trigger is extraterritorial in a way that strains the concept of state law. As practitioners at Reed Smith note, the obligation attaches to any ad that could reach a New York audience — which, for any national digital campaign, means every ad. A small advertiser in Ohio running a generic AI-generated spot on Instagram is now subject to a New York disclosure mandate it may never have heard of, with no intentional targeting of the state. That is not how consumer-protection statutes have historically worked, and it raises the federalism question of one state effectively writing national advertising rules. It also invites the familiar patchwork problem: if a dozen states each pass a differently worded version, advertisers comply with the strictest and the others become dead letters.
Second, the key terms are undefined where it counts. The statute requires a "conspicuous" disclosure but never says what conspicuous means across formats — a six-second vertical video, a banner, a carousel. Cooley's analysis flags that the "actual knowledge" standard, while a welcome guardrail against strict liability, creates its own ambiguity: advertisers will need documented internal escalation practices to prove what they did and did not know. And because the definition reaches any "software algorithm," not just generative AI, it arguably sweeps in long-standing CGI and retouching techniques that no one thinks of as deception.
Third, enforcement is underspecified. The statute sets penalties but, as the law-firm analyses observe, does not clearly say how violations are counted — per ad, per impression, per campaign — which turns a $5,000 repeat penalty into open-ended exposure for a national advertiser running thousands of placements. Predictable liability is a precondition for good-faith compliance; this law does not yet provide it.
A better template
None of this argues against disclosure. It argues for tighter drafting. A workable version would: define "conspicuous" with format-specific safe harbors, as the FTC does in guidance; tie jurisdiction to ads targeted at New York rather than any ad that could theoretically reach it; confirm enforcement rests solely with the Attorney General with no private right of action; and specify per-campaign rather than per-impression violation counting. Those changes would preserve every consumer-protection benefit while removing the traps.
New York has the right instinct: in an influencer economy increasingly populated by synthetic faces, a viewer should be able to tell person from render. The instinct deserves a statute built to be obeyed, not one that turns every national advertiser into an inadvertent first offender.