New York Just Made AI Influencers a Compliance Problem

New York's synthetic-performer law is now active, and the useful read for creator marketing is operational. An AI avatar in a campaign now carries a compliance review alongside creative approval. If the ad reaches New York and the advertiser knows the human-looking performer is synthetic, the campaign needs a disclosure decision that survives production, trafficking, platform upload, and later review.

The law took effect June 9, 2026, 180 days after Gov. Kathy Hochul signed S.8420-A/A.8887-B in December. AP reported that Hochul's office calls it a first-in-the-nation law and that the rule applies to advertisements in any medium. The state announcement ties the rule directly to AI-generated performers appearing in social media and digital advertising, which is why this matters to brands using virtual creators, AI product demos, synthetic testimonials, or human-looking background talent in paid social.

What counts as the performer

The definition is specific. A synthetic performer is a digital asset created, reproduced, or modified by computer, using generative AI or a software algorithm, and intended to give the impression that it is performing as a human who is not recognizable as an identifiable natural person. That catches more than prompt-generated avatars. Reed Smith and Kelley Drye both flagged that the software-algorithm language may reach traditional CGI-style human assets, while the intent element can matter for mannequins, cartoons, abstract avatars, or other images never meant to read as real people.

The identifiable-person boundary is just as important. A synthetic version of a real celebrity, dead performer, employee, or creator clone sits closer to publicity-rights and consent law than to this disclosure provision. Skadden's summary separates the synthetic-performer ad rule from New York's companion expansion of post-mortem publicity protections. A brand using a fictional AI influencer has a disclosure question; a brand using the likeness of a real person has a consent and rights question before it ever gets to the label.

The disclosure has to be built into the asset

The statute requires a conspicuous disclosure in the advertisement, but it does not prescribe the words, size, placement, or duration. That blank space is the workflow problem. Skadden wrote that the law does not specify the form of the disclosure. Kelley Drye noted that some advertisers expect to use the statutory term while others may choose more consumer-readable language such as AI-generated image. Until regulators create examples, the safer operating standard is a label a reasonable viewer can notice and understand in the same feed where the ad appears.

That means the disclosure cannot live only in a legal note, agency deck, or creator brief. The statute says the advertisement itself must disclose that a synthetic performer appears in it. For a short-form social ad, the file should answer the same questions a reviewer would ask after the fact: whether the label is in the video, caption, platform disclosure field, landing page, or some combination; whether viewers see it before the synthetic performer does commercial work; and whether the label remains visible after platform cropping, caption truncation, or format changes.

The exemption list is where briefs will break

The law's carve-outs are narrow. Promotional materials for expressive works, including films, TV programs, streaming content, documentaries, video games, and similar audiovisual works, are outside the rule when the synthetic performer's use in the ad is consistent with its use in the underlying work. Audio-only ads and AI used solely for language translation of a human performer are also outside the disclosure requirement.

Those carve-outs leave the ordinary creator-marketing use case exposed to analysis. A trailer for a game character may be different from an AI fashion model selling a jacket, an AI spokesperson reading a brand claim, or a synthetic background crowd in a retail ad. Kelley Drye pointed out unresolved questions around background performers and body parts such as a hand modeling a watch because the statute does not distinguish principal talent from extras. The practical answer is to classify the asset before trafficking, then save the classification with the creative approval.

Distribution makes this a national workflow

Lowenstein Sandler's Reuters piece gives the campaign-planning implication: the rule applies to advertisements distributed to New York audiences, even when the advertiser is outside the state. Reed Smith put the same point more broadly for ads that may reach New Yorkers. A national paid-social buy therefore needs a New York answer unless the brand geo-fences the campaign with confidence or treats the New York disclosure standard as the default.

Publishers and platforms become the second-line compliance surface. The statute says media used for advertising are generally outside the rule unless they receive written notice that an ad contains an undisclosed synthetic performer and then get no more than five days, or as soon as technically or practically feasible, to stop distribution or insert the disclosure. It also bars entities that publish or disseminate covered ads from removing or altering the disclosure before publication. For a brand, that turns the delivery spec into part of the compliance file: the label has to be supplied in a form the publisher or platform will preserve.

The penalty amounts are modest compared with a national media budget, but that is the wrong place to price the risk. The larger cost is the process change: synthetic performers need an asset-level disclosure record before the campaign launches. Brands that already run paid creator programs can borrow the same discipline they use for endorsement disclosures: identify the commercial relationship, decide where the viewer will see the label, make the label a condition of approval, and keep the proof. New York has made that file necessary for a new class of artificial talent.