What is New York's synthetic performer disclosure law?

A synthetic performer is an AI-made person in an advertisement: an avatar presenter, a realistic AI model wearing your clothes, an AI voice that reads your script. New York and California now ask advertisers to say so. This guide reads both statutes, shows where they differ, and covers the EU parallel and what marketplaces ask of sellers. It is general information, not legal advice, and we checked every source on October 21, 2026.

The statutes put the duty on the advertiser, not on the tool that made the performer. For the wider map, start from our AI disclosure laws hub and the United States page.

What New York's law says

The law is General Business Law section 396-b. The Assembly's bill page shows it signed on December 11, 2025 as Chapter 617, and the act takes effect on the 180th day after it became law, which is June 9, 2026. Subdivision 3 holds the duty: "Any person engaged in the business of dealing in any property or service who for any commercial purpose produces or creates an advertisement respecting any such property or service, in any medium or media in which such advertisement appears, shall conspicuously disclose in such advertisement that a synthetic performer is in such advertisement, where such person has actual knowledge. A violation of this subdivision shall result in a civil penalty of one thousand dollars for a first violation, and five thousand dollars for any subsequent violation."

• What counts. A synthetic performer is "a digitally created asset created, reproduced, or modified by computer, using generative artificial intelligence or a software algorithm, that is intended to create the impression that the asset is engaging in an audiovisual and/or visual performance of a human performer who is not recognizable as any identifiable natural performer." Note "or a software algorithm", and "and/or visual": read literally, a still image of a synthetic human counts as well as a video avatar.
• Who is bound. The advertiser side: the person in the business of dealing in property or services who produces or creates the ad. Media that publish or disseminate an ad are outside the law (subdivision 8), and subdivision 6 leaves the protections of 47 U.S.C. section 230 unchanged. Whether an agency or a freelance producer is "engaged in the business of dealing in any property or service" is not settled by the text.
• Knowledge. The duty applies "where such person has actual knowledge". The statute does not say knowledge of what. The natural reading is knowledge that the ad contains a synthetic performer, so an advertiser who generated the performer is caught, and one who bought a stock clip without knowing it was synthetic may not be.
• What to say. The statute prescribes no wording, size, duration or placement beyond "conspicuously" and "in such advertisement".
• Exceptions. Ads and promotional materials for expressive works such as films, television, streaming content, documentaries and video games, if the use matches the work (subdivision 4). Audio advertisements and cases where AI "solely involves the language translation of a human performer" (subdivision 7).
• Real people. A real, identifiable person's likeness is outside the definition. Subdivision 5 leaves rights under sections 50, 50-f and 51 of the Civil Rights Law untouched. A disclosure does not replace consent.
• Enforcement. The statute names no enforcer, no cure period, no private right of action and no per-day rule. Who sues is therefore open. We found no enforcement action under it, and subdivision 3, unlike subdivision 2 of the same section, does not say whether it reaches ads shown only outside New York.

Do not read the law as a general deepfake label rule. The enacted text covers advertisements only.

What California's SB 1050 adds

California SB 1050 was approved on September 16, 2026 (Chapter 246) and adds section 17610 to the Business and Professions Code. The bill prints no operative date, so by our derivation from the general rule for California statutes it takes effect on January 1, 2027. Subdivision (b) says: "It is unlawful for any person to create and cause to be published in an advertising medium an advertisement that prominently includes a synthetic performer without a clear and conspicuous disclosure that the advertisement includes a synthetic performer."

• Prescribed wording. Subdivision (c): "A disclosure required by this section shall use wording substantially similar to “this performance features a synthetic performer” or “no human performer is depicted.”"
• "Prominently." The duty applies only if the performer is in the foreground demonstrating or illustrating the product, provides or voices the narration or message, or illustrates or reacts to it. A background extra is outside.
• "Clear and conspicuous." The statute defines it as "a disclosure that is difficult to miss, easily understandable, and presented in a manner that a reasonable consumer would notice, read, and comprehend, taking into account the medium, format, and context in which the advertisement appears."
• What counts. "a digital figure, voice, or representation created in whole or in part using generative artificial intelligence" that creates "the realistic impression of the audio, audiovisual, or visual performance of a human performer who is not recognizable as any identifiable natural person." Unlike New York, audio-only ads are included and the AI must be generative.
• Exceptions. Ads for expressive works that match the work, and ads where generative AI "solely involves the language translation of a human performer or other accessibility features."
• Enforcement. A violation "constitutes a violation of Section 17500 and may be enforced pursuant to Chapter 5 (commencing with Section 17200)", so the False Advertising Law and the Unfair Competition Law apply. Section 17500 provides that a violation is a misdemeanor punishable by up to six months in county jail, a fine up to $2,500, or both, and the Legislative Counsel's digest notes that the bill expands the scope of a crime. How that applies to a missing disclosure is untested. The statute also bars an advertising medium from carrying an ad after a court order finds a violation and the medium is served with it.

Does your ad have a synthetic performer?

These are our readings of the definitions, not rulings. Several are open.

• An AI avatar presenter reading your script. Yes in both states, if you know it is synthetic (New York) and it is prominent (California).
• A realistic AI-generated model wearing your product in a photo ad. Likely yes. New York's text says "visual performance", and California's covers a "visual" performance.
• A stock AI avatar clip you bought without knowing it was AI. New York's "actual knowledge" element may protect you. California's text states no knowledge element.
• A cartoon mascot or an obviously stylized character. Probably not. Both definitions turn on the impression of a human performer, and California's says "realistic impression". Nobody has tested it.
• An AI voice-over on a video. California names a "voice". New York exempts audio-only ads but a video is not one, and whether a synthetic voice alone is a synthetic performer there is open.
• An AI likeness of a real, identifiable person. Outside both definitions, so these laws do not govern it. Consent and publicity laws do.
• AI used only to translate a real performer's speech. Exempt in both.

What the disclosure can say, and where it goes

Both statutes ask for a disclosure in the ad. These suggestions are ours, not a legal standard.

• One wording for both states: "This performance features a synthetic performer." It is the California phrase, and New York prescribes none. Ask a lawyer before relying on any single wording.
• In the ad itself, because New York says "in such advertisement". A line on the landing page is not the ad. Repeat it in every crop and length you run.
• Conspicuous. Neither statute gives a size, a duration or a contrast ratio. California's definition, "difficult to miss", is the closest test. On video, show it long enough to read, and say it or caption it.
• Accessible. On-screen text is invisible to a screen reader unless there is a text equivalent, and a spoken line needs captions. Our guide to accessible AI labels covers the pixels problem and the contrast criteria.

Our free accessible AI label generator has synthetic performer presets for New York and California that write wording and markup with a text equivalent. The AI disclosure laws finder shows which rules may reach your ads.

What platforms and marketplaces ask

• Amazon. Amazon is reported to require sellers to add the keyword contains-synthetic-performer to the XMP dc:subject field of images and videos, A+ Content included, that show a photorealistic person created entirely by AI, and to show a disclosure on listings "where applicable". The reports, including eWeek's summary of CNBC's July 23, 2026 story, link it to New York's law. We could not read Amazon's own page, which requires a seller login. Reports differ on whether images uploaded earlier must be re-tagged, so we make no claim about that.
• Google Ads. Google's AI label setting, introduced in July 2026, shows visible overlays for campaigns targeting the EU, India and New York, and Google says the setting does not guarantee compliance. See its policy update.
• TikTok ads. Advertisers tick "This ad contains AI-generated content" for fully AI-generated or significantly AI-modified image, video or audio, per TikTok's Ads Manager help.

These are platform rules and platform tools. The statutes ask the advertiser for a disclosure in the ad, and a keyword in a file is not that. How to write the keyword and why it may not survive resizing is in our guide to AI image metadata.

The EU parallel

Article 50(4) of the EU AI Act asks deployers of AI systems that generate or manipulate image, audio or video content that is a deep fake to disclose that it was artificially generated or manipulated. The Commission's Guidelines (C(2026) 5054 final, July 20, 2026, non-binding) say in paragraph 113 that "persons" include "realistic AI-generated human avatars or personas". Paragraph 116 lists "AI-generated video featuring an AI-generated depiction of a celebrity influencer in an advertising or promotional context" as a deep fake. The examples after paragraph 124 include "a realistic synthetic influencer testing out a sponsored real product" among deep fakes that are not creative works and so get no lighter disclosure regime. Paragraph 122 adds that some advertisements might count as evidently creative or fictional works in specific situations and not in others.

The EU rule differs from both US laws. It has no actual knowledge element, covers any deep fake and not only performers, and prescribes no wording. It binds deployers in the EU and, under paragraph 13, deployers elsewhere who foresee that their output will be used in the Union. A US brand running ads to EU audiences should read our Article 50 page.