Client Alert: New York’s Synthetic Performer Disclosure Law is Now in Effect
August 31, 2026
New York’s synthetic performer disclosure law took effect on June 9, 2026.[1] Codified at New York General Business Law § 396-b, it is the first state law in the nation directed at the use of digitally created human performers in commercial advertising. What makes it notable is that the obligation is triggered by the presence of a defined asset in an advertisement, not by whether the advertisement is otherwise false or misleading.
Under § 396-b(3), a person engaged in the business of dealing in any property or service who, for any commercial purpose, produces or creates an advertisement respecting that property or service, in any medium or media in which the advertisement appears, must conspicuously disclose that a “synthetic performer” is in the advertisement, where that person has actual knowledge. Each element matters, and each is discussed below.
Violations carry a civil penalty of $1,000 for a first violation and $5,000 for any subsequent violation. The statute does not define the unit of violation, that is, it does not say whether penalties accrue per advertisement, per placement, per day, or otherwise. This means that cumulative exposure in a high-volume campaign cannot be modeled with confidence. The enacted text also identifies no enforcement authority and creates no express private right of action. As a practical matter, enforcement is expected to rest with the New York Attorney General under its general authority, including Executive Law § 63(12). Note as well that § 396-b(2), a provision predating this amendment, continues to make certain false statements in advertising a misdemeanor; the section is not exclusively civil.
Because the law is already in effect, live creative, not just future campaigns, should be the immediate priority.
Who Is Covered, and Where
Two threshold questions are unsettled and worth flagging to your business team before you build a compliance program around either answer.
Brand or agency? Subdivision 3 reaches a person engaged in the business of dealing in any property or service who produces or creates an advertisement respecting “any such” property or service. Read literally, that describes the advertiser whose goods or services are being promoted — not a pure agency or production company, whose own service is advertising rather than the subject of the advertisement. We think the better reading places primary responsibility on the brand. Agencies and production companies should not treat that as a safe harbor, however: the allocation is contractual in practice, and an in-house or affiliated production function will often sit on the brand’s side of the line.
Does it reach out-of-state campaigns? The statute contains no express extraterritoriality or “could reach New York” test. Subdivision 2 applies to advertisements placed before the public “in this state”; subdivision 3 conspicuously omits that phrase. Much of the published commentary nonetheless assumes that any advertisement reaching a New York audience is covered regardless of where the advertiser sits, and you should expect to hear that position from other counsel. The conservative operating assumption is to treat nationally distributed and New York-targeted creative as in scope, while recognizing that the question is genuinely open and that a categorical “every campaign everywhere” reading is not supported by the text.
What Qualifies as a “Synthetic Performer”
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, 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.”[2]
“Realistic fake human” is useful shorthand for explaining this to a marketing team, but it is not the statutory test, and neither is “realism”, that word does not appear in the statute. The elements are: (i) a digitally created asset; (ii) created, reproduced, or modified by computer using generative AI or a software algorithm; (iii) intended to create the impression that the asset is engaging in an audiovisual and/or visual performance; (iv) of a human performer; and (v) one who is not recognizable as any identifiable natural performer. Edge cases should be worked through against those elements rather than against a general impression of how lifelike the asset looks.
Two points follow that often surprise clients. First, the obligation is not triggered by the use of AI in advertising generally; an AI-drafted headline, an AI-generated landscape, or an AI-assisted edit does not by itself require disclosure. Second, the statute is not limited to generative AI: it also reaches assets created, reproduced, or modified by computer using “a software algorithm.” Whether conventional CGI pipelines, motion capture, or digital compositing fall within that phrase has not been tested. Motion capture presents an additional wrinkle, because the resulting performance derives from a real performer and may remain recognizable as an identifiable natural performer, which would take it outside the definition.
Common candidates for coverage include AI-generated models in product photography, synthetic “customer” spokespeople, and digitally created background figures in lifestyle imagery. The statute does not distinguish featured performers from background “extras,” and it does not address partial depictions, an AI-generated hand modeling a watch, for instance. These are unresolved. Treating them as covered is the cautious course, but it is a risk posture rather than a statutory requirement, and it should be presented to the business as a choice.
Actual knowledge. Actual knowledge is an express condition of the duty. The statute does not say constructive knowledge or “should have known” suffices. Worth noting: the bill as originally introduced extended the trigger to knowledge held by an “agent or employee” of the covered person; the enacted version dropped that language.[3] How knowledge is attributed within a corporate advertiser is therefore unresolved. This creates a real tension with the documentation practices recommended below; thorough provenance records can supply the very knowledge that triggers the obligation. We nonetheless recommend documentation, because deliberate ignorance is a poor litigating posture, because the same records are needed to apply the exemptions, and because the adjacent regimes discussed below impose obligations that do not turn on knowledge at all.
Express Exclusions and Scope Limitations
The statute contains several express exclusions. Note that subdivisions 4, 7, and 8 each provide that “this section” shall not apply — not merely the disclosure subdivision — so where they apply, they take the conduct outside § 396-b altogether. Other categories that clients often assume are exempt are, in fact, definitional questions requiring fact-specific analysis.
Identifiable real performers (definitional, not an exemption). The definition reaches only a performance of a human performer “who is not recognizable as any identifiable natural performer.” That language places many advertisements featuring recognizable real people outside the definition, and routine retouching, color correction, or AI-assisted enhancement will typically not change that. It is not, however, a blanket exemption for every use of a real person: an asset modified to the point that the underlying individual is no longer recognizable may fall back within the definition. Separately, and importantly, § 396-b(5) expressly preserves all rights under Civil Rights Law §§ 50, 50-f, and 51 and under any other law. Digital replicas of real people, living or deceased, raise right-of-publicity and privacy exposure that this disclosure statute neither displaces nor resolves, and New York amended its deceased-performer digital replica provisions in a companion bill signed the same day. A disclosure does not cure a consent problem.
Non-human and stylized imagery (definitional, not an exemption). The statute does not list cartoons, animated mascots, robots, mannequins, product renderings, or stylized avatars as exempt. An asset not intended to create the impression of a human performance falls outside the definition on its own terms, which will cover most clearly non-human imagery. But the analysis runs through the statutory elements, not through a category list: a stylized digital human intended to read as a person can qualify.
Audio advertisements. Section 396-b(7)(a) excludes audio advertisements, and the definition is independently limited to “audiovisual and/or visual” performance. (The sponsor’s memorandum described a definition covering “audio, audiovisual, and/or visual” performance; the enacted text dropped “audio,” which supports the narrower reading.) The harder case, which the statute does not resolve, is a synthetic voice paired with visual content. A radio or podcast spot is clearly excluded. An AI voice over an audiovisual advertisement should not be assumed to be.
Language translation — read this one narrowly. Section 396-b(7)(b) excludes advertisements where the use of artificial intelligence “solely involves the language translation of a human performer.” The word “solely” does the work. Straight dubbing or subtitling of a real performer’s speech fits. AI lip-sync or visual mouth reanimation to match a translated track does not obviously fit, because it modifies the visual performance rather than translating language, and localization workflows that also generate or synthesize other elements will fall outside the carve-out entirely. This is a high-volume workflow for many advertisers and, in our view, the provision most likely to be misapplied.
Promotional materials for expressive works. Section 396-b(4) provides that the section does not apply to advertisements and promotional materials for expressive works, including motion pictures, television programs, streaming content, documentaries, video games, and similar audiovisual works, provided that the use of the synthetic performer in the advertisement is consistent with its use in the expressive work. The condition is about consistency of use, not merely whether the same character appears in both. A synthetic character carried from a film into its trailer in the same role fits comfortably. The same character repurposed into a different promotional context, or used in ways it is not used in the underlying work, may not.
Media, publishers, and platforms. Section 396-b(8) provides that the section does not apply to any medium used for advertising — newspapers, magazines, television networks and stations, streaming services, cable systems, billboards, and transit advertising — by whom a non-compliant advertisement is published or disseminated. As enacted, this is an unqualified exclusion.[4] Two caveats. It protects the medium acting in that publication or dissemination role; it is not a shield for an entity that also produces or creates the advertisement or otherwise falls within § 396-b(3), which matters for platforms and retail media networks running their own creative. And § 396-b(6) separately provides that nothing in the section limits or enlarges the protections that 47 U.S.C. § 230 confers on an interactive computer service for content provided by another information content provider.
What a Compliant Disclosure Should Say
The statute requires a “conspicuous” disclosure but does not define that term, prescribe wording, or set placement or formatting requirements. Until the New York Attorney General or a court provides interpretive guidance, no particular formulation is a safe harbor.
The most conservative approach tracks the statutory language, since the obligation is to disclose that a synthetic performer is in the advertisement: “This advertisement contains one or more synthetic performers generated using artificial intelligence.” Plainer formulations that most consumers will parse more readily include “AI-generated performer”; “This ad includes an AI-created person”; and “The performers shown in this ad are computer-generated and are not real people.” Where the medium allows, pairing the plain-language version with the statutory term is the belt-and-suspenders option.
For placement and prominence, the Federal Trade Commission’s “clear and conspicuous” standard is a practical benchmark: the disclosure should appear close to the relevant content; be sufficiently prominent in size, contrast, and duration to be noticed and understood; and not be buried in fine print or behind a hyperlink. For campaigns directed at non-English-speaking New York audiences, the disclosure should appear in the language of the advertisement.
Adjacent Obligations That This Statute Does Not Address
Compliance with § 396-b does not resolve the following, and in at least one case a disclosure will not cure the underlying problem.
Synthetic testimonials are a separate and larger problem. The FTC’s Rule on the Use of Consumer Reviews and Testimonials prohibits fabricated consumer testimonials outright, with civil penalty exposure well beyond the amounts in § 396-b.[5] An AI-generated “customer” endorsing a product is not made lawful by labeling the performer as synthetic; the endorsement itself is the violation. Section 5 of the FTC Act and state UDAP statutes apply independently. Clients should be told plainly that this is not a disclosure problem.
Right of publicity and digital replicas. As noted above, § 396-b(5) preserves Civil Rights Law §§ 50, 50-f, and 51. Any use of a real person’s likeness, living or deceased, replicated or enhanced, should be run through a consent and clearance analysis separate from the disclosure question.
EU AI Act transparency obligations. For advertisers with European distribution, Article 50 of the EU AI Act became applicable on August 2, 2026 and imposes its own marking and labeling requirements for AI-generated content and deepfakes.[6] The obligations are not coextensive with New York’s, and a disclosure designed for one will not necessarily satisfy the other.
SAG-AFTRA commitments. Signatory advertisers and agencies should check the current SAG-AFTRA Commercials Contract before deploying synthetic performers. Its provisions on synthetic performer use are contractual obligations that operate independently of, and in some respects more restrictively than, the statute.
Preemption and Litigation Watch
Two sources of uncertainty bear on how much to invest in compliance now.
First, hours after the Governor signed this bill on December 11, 2025, the White House issued an executive order seeking to curtail state-level AI regulation, directing the Department of Justice to challenge state AI laws and conditioning certain federal broadband and discretionary funding on states not maintaining “onerous” AI laws. Whether a likeness-disclosure requirement of this kind falls within the order’s intended scope is contested, and the order itself faces legal challenges. Second, a disclosure mandate of this type is a natural candidate for a compelled-commercial-speech challenge under the First Amendment, particularly given the absence of any defined “conspicuous” standard. The act includes a severability clause, so a successful challenge to one provision would not necessarily take down the rest.
As of the date of this alert, we are aware of no major litigation directly challenging § 396-b and no interpretive guidance from the New York Attorney General. Our recommendation is to comply now and monitor: the compliance cost is modest relative to the exposure, and the adjacent obligations described above would survive even if this statute did not.
Recommended Steps
The statute does not prescribe an internal compliance program. The following are practical risk-management measures, not statutory requirements, and each should be scaled to how heavily your organization uses synthetic imagery.
Audit live creative first. Review currently running advertisements and e-commerce product listings for AI- or software-generated human likenesses, including AI models, synthetic spokespeople, and digitally created background figures. Because the law is in effect, live creative is the priority. Flag any synthetic testimonials for separate escalation under the FTC rule.
Apply disclosures where the elements are met. Where a covered person has actual knowledge that an advertisement includes a synthetic performer and no exclusion applies, add a conspicuous disclosure using the wording and placement guidance above.
Document the creative process — deliberately. Retain creative briefs, generation prompts, vendor representations, and records of the tools used. Read the actual-knowledge discussion above before designing this: the goal is a record that lets you apply the exemptions and demonstrate good-faith process, not an undifferentiated archive.
Add a creative review checkpoint. Insert a written step in the creative approval workflow addressing (1) whether the advertisement contains an asset that may meet the synthetic performer definition, (2) whether any exclusion applies, and (3) whether a disclosure has been included. Catching this pre-publication is materially cheaper than responding to an inquiry.
Update agency and vendor agreements. Address identification of AI- or software-generated human likenesses in deliverables; allocation of responsibility for applying and maintaining disclosures across placements and channels; preservation of provenance information; notice obligations; and indemnification for failures to disclose. Given the open brand-versus-agency scope question, the contractual allocation is doing more work here than the statute is.
Update influencer and creator programs. Brands relying on influencer or user-generated content should address disclosure obligations in program policies and creator agreements, including training, pre-approval of content created with generative tools, and audit and remediation rights for non-compliant posts.
Contact Us
Please contact us if you have questions about whether your advertising content falls within the scope of this law, how to design compliant disclosures, how to update your agency and vendor agreements, or how to integrate these requirements into your existing advertising compliance program.
For questions about this alert, please contact:
Cameron Robinson | Partner | crobinson@crokefairchild.com | +1.872.224.2920
Bakhtawar Mirjat | Associate | bmirjat@crokefairchild.com | +1.872.224.2952
References
[1]S.8420-A / A.8887-B, signed by Governor Hochul on December 11, 2025 as Chapter 617 of the Laws of 2025. The act took effect on the 180th day after signing, i.e., June 9, 2026. The new provisions are codified at N.Y. Gen. Bus. Law § 396-b(1) and (3)–(8).
[2]N.Y. Gen. Bus. Law § 396-b(1)(c). The statute separately defines “artificial intelligence” and “generative artificial intelligence” at § 396-b(1)(a)–(b).
[3]Compare the original print of S.8420, § 396-b(3) (“where such person, or agent or employee thereof, has actual knowledge”) with the enacted S.8420-A, § 396-b(3) (“where such person has actual knowledge”).
[4]The original print of S.8420 conditioned the media exclusion on a five-day cure period following written notice, and included a companion anti-tampering provision. Neither was enacted. Several published summaries of this law continue to describe the superseded text.
[5]Rule on the Use of Consumer Reviews and Testimonials, 16 C.F.R. pt. 465.
[6]Regulation (EU) 2024/1689, art. 50. The transparency obligations became applicable on August 2, 2026.