What is New York's synthetic performer disclosure law and when did it take effect?

New York's synthetic performer disclosure law, codified at N.Y. General Business Law section 396-b, is a first-in-the-nation measure requiring advertisers to conspicuously disclose when an advertisement contains an AI-generated synthetic performer, and its disclosure requirements took effect on June 9, 2026, after the law was signed at the end of 2025. Its defining feature, compared with other AI-transparency measures, is that it places the disclosure obligation directly on advertisers — the businesses that create or produce advertisements — rather than on the AI tools used to make the content. The law applies to any person engaged in the business of dealing in property or services who creates or produces an advertisement for a commercial purpose, and the duty is triggered where that person has actual knowledge that a synthetic performer appears in the advertisement. When those conditions are met, the advertisement must conspicuously disclose that it contains a synthetic performer, so an ordinary viewer can tell that the performing figure they are watching is AI-generated rather than a real person. This advertiser-level approach is what distinguishes it from provenance-focused regimes. Laws and standards that require AI systems to embed watermarks, maintain provenance metadata, or offer detection tools operate at the level of the AI tool or model; New York's law operates at the level of the ad and the advertiser, requiring a visible disclosure in the creative itself. For advertisers, that means it is a concrete compliance task attached to their own campaigns, with defined civil penalties for non-compliance, rather than a background technical standard handled by their vendors. The law is also deliberately targeted. It concerns 'synthetic performers' — AI-generated fictional humans who appear to perform in an ad and are not real, identifiable people — rather than every use of AI in advertising, and it carries specific exceptions, including for audio-only ads and for AI used solely to translate a human performer's language. Because it is a new, first-of-its-kind measure that took effect in mid-2026, and because similar state proposals are emerging, advertisers should treat it as the leading edge of a broader trend toward requiring disclosure of synthetic performers, and should confirm the current requirements against official New York sources. For the converging California regime see the California AI Transparency Act guide , and track state developments on the Policy Change Tracker . The organizing principle is that section 396-b requires advertisers to conspicuously disclose AI-generated synthetic performers in their ads, effective June 9, 2026, placing the duty on advertisers rather than AI tools.

What exactly counts as a 'synthetic performer' under the law?

A 'synthetic performer' under New York's law is a digitally created asset — created, reproduced or modified by computer using generative AI or a software algorithm — that is intended to give the impression it is engaging in an audiovisual or visual performance of a human performer who is not recognizable as any identifiable natural person; in plain terms, it is an AI-generated fictional 'human' who appears to act or perform in an advertisement but is not a real, identifiable person. This precise definition is narrower than 'any AI-generated content', and understanding its elements is essential to knowing when the disclosure duty applies. There are four elements working together. First, the asset must be digitally created — created, reproduced or modified by computer — which captures computer-generated and AI-modified imagery rather than ordinary filmed footage of real people. Second, it must be made using generative AI or a software algorithm, tying it specifically to AI or algorithmic generation. Third, and centrally, it must be intended to give the impression that it is engaging in a visual or audiovisual performance of a human — that is, it must appear to be a person performing or acting on screen, not merely a static graphic or an inanimate element. Fourth, that apparent human must not be recognizable as any identifiable natural person, meaning it is a fictional AI 'human' rather than a depiction of a real, named individual. The combined effect deliberately excludes two adjacent categories that advertisers should not confuse with synthetic performers. It does not cover AI depictions of real, identifiable people — for example an AI-generated likeness of a celebrity or a specific real individual — because those raise separate legal issues around right of publicity, likeness and consent, addressed by other laws rather than this disclosure requirement. And it does not cover AI uses that do not create a performing human at all: AI-generated backgrounds, scenery, products, logos, graphics, music or text are not synthetic performers, because none of them is a 'human' appearing to perform. The performance-of-a-human element is the crux. This matters practically because it prevents both under-disclosure and over-disclosure. An advertiser must disclose when it features an AI-generated fictional presenter, spokesperson-like figure, or actor who appears human but is not a real person; it need not attach a synthetic-performer disclosure to an ad merely because AI was used to generate a background or retouch a product image. Drawing the line correctly — is there an AI-generated 'human' who appears to perform, and are they not a real identifiable person? — is the key judgement. For definitions of the surrounding terminology see the compliance glossary , and for how likeness of real people is treated see the FTC AI endorsement guide . The organizing principle is that a synthetic performer is an AI-generated fictional human who appears to perform in an ad and is not a real, identifiable person, which excludes real-person likenesses and non-performing AI elements.

Who has to comply, and what does the 'actual knowledge' requirement mean?

The law requires compliance from any person engaged in the business of dealing in property or services who creates or produces an advertisement for a commercial purpose, and the disclosure duty is triggered where that person has actual knowledge that a synthetic performer appears in the advertisement — so the obligation falls on the commercial advertiser or producer of the ad, and it attaches to what that advertiser actually knows about its own creative. Both parts of this — who is covered and the knowledge standard — shape how the law works in practice. On who is covered: the duty targets the business that creates or produces the advertisement for a commercial purpose, meaning the advertiser or its producer, not the platform or publisher that merely displays the ad. A social platform showing an ad is not the party creating it for a commercial purpose in the relevant sense; the advertiser that made the creative is. This places responsibility where the knowledge and control over the creative actually sit, which is sensible because the advertiser is the party that decided to use a synthetic performer and knows how the ad was made. On the actual-knowledge standard: the obligation to disclose arises where the covered person has actual knowledge that a synthetic performer is in the advertisement. 'Actual knowledge' is a meaningful qualifier — it ties the duty to what the advertiser genuinely knows, rather than imposing liability for synthetic performers the advertiser had no awareness of. In practice, however, this is not a loophole for deliberate users of the technology, because an advertiser that chooses to create an ad with an AI-generated synthetic performer will ordinarily know that it did so; the synthetic performer is a deliberate creative decision, not something that appears by accident. So for the typical case the law is aimed at — an advertiser intentionally using an AI 'human' presenter — the actual-knowledge condition will be satisfied, and the disclosure duty applies. Where the standard matters more is at the margins, such as content produced by third parties whose AI use the advertiser genuinely did not know about; but the prudent course there is to ask and to establish knowledge rather than to rely on ignorance, especially since building a production check is easy. When both conditions are met, the required disclosure must be conspicuous — noticeable to the audience rather than hidden — consistent with how disclosure standards operate elsewhere in advertising law. The practical implication is that advertisers running visual or audiovisual campaigns to New York audiences should assume the law applies whenever they knowingly use an AI synthetic performer, and should build a workflow step to identify such uses and attach a clear disclosure. For related disclosure-standard principles see the FTC AI endorsement guide , and screen disclosures with the Keyword Risk Checker . The organizing principle is that the duty falls on the commercial advertiser or producer of the ad and is triggered by actual knowledge of a synthetic performer, which deliberate users will ordinarily have.

What are the exceptions and penalties?

New York's synthetic performer disclosure law contains specific exceptions that narrow its reach — most notably for audio-only advertisements and for cases where AI is used solely for language translation of a human performer — and it carries defined civil penalties of one thousand dollars for a first violation and five thousand dollars for subsequent violations, giving the disclosure requirement real consequences while remaining targeted in scope. Taking the exceptions first: the audio-only exception reflects that the law is concerned with visual or audiovisual synthetic performers — AI-generated 'humans' who appear to perform on screen — so an advertisement that is audio-only, with no visual performance, falls outside the requirement. This makes sense because a synthetic performer, as defined, gives the impression of a visual or audiovisual performance; a purely audio ad does not present a visual performing figure of the kind the law addresses. The translation exception addresses a common and benign use of AI: where a real human performer's genuine performance is simply translated into another language using AI — changing the language of the speech while the performer remains a real person — this does not create a synthetic performer, because there is a real, identifiable human at the core and only the language has been altered. The exception prevents the law from sweeping in legitimate localisation of real performances. These exceptions help advertisers judge when the duty does not apply, but they should be read narrowly: the audio-only exception does not help a video ad, and the translation exception applies only where AI is used solely for translation of a genuine human performer, not where AI generates or substantially creates a performing figure. On penalties, the structure escalates: a first violation carries a civil penalty of one thousand dollars, and subsequent violations carry a civil penalty of five thousand dollars each. Per instance these figures are modest, but they can accumulate across multiple ads or campaigns, and the escalation to five thousand dollars for repeat violations creates a clear incentive to fix the problem systematically rather than treating individual penalties as a cost of doing business. The practical message the penalty structure sends is to build disclosure into the production process so that violations do not recur, because the law specifically increases the cost of repeated non-compliance. For advertisers, the combination of targeted exceptions and escalating penalties means the sensible approach is to apply the disclosure wherever a visual synthetic performer is knowingly used, rely on the exceptions only where they clearly fit, and institutionalise the practice to avoid the higher repeat-violation penalty. Track how comparable measures evolve on the Policy Change Tracker . The organizing principle is that the law excepts audio-only ads and AI-translation-only uses and imposes civil penalties of one thousand dollars for a first violation and five thousand dollars for subsequent ones, incentivising systematic compliance.

How should advertisers build compliance into their creative process?

Advertisers should build compliance with New York's synthetic performer law into their creative process by treating it as a repeatable production checkpoint rather than a legal puzzle: identify in production whether any creative contains an AI-generated synthetic performer, add a conspicuous disclosure where it does, avoid over-disclosing on AI uses that are not synthetic performers, and coordinate the disclosure with platform AI labels and tool-level provenance — a routine that, once embedded, makes compliance automatic. The reason a process focus works is that the legal test is simple but easy to overlook in a fast production pipeline. The substantive question — is there an AI-generated 'human' who appears to perform, and are they not a real identifiable person? — can be answered by whoever makes the creative, because they know how it was produced. The risk is not misunderstanding the law but failing to check consistently. So the first step is to insert a synthetic-performer check into the creative workflow: whenever AI is used to generate a performing human figure, flag it. This can be built into creative briefs, production notes and the review stage, so that the question is asked on every visual or audiovisual asset rather than remembered ad hoc. The second step is the disclosure itself: where a synthetic performer appears in a visual or audiovisual ad reaching New York audiences, add a clear, noticeable disclosure that the ad contains a synthetic performer. Conspicuousness matters — the disclosure should be positioned and presented so an ordinary viewer notices it, not buried where it will be missed. The third step is discrimination in the other direction: avoid attaching synthetic-performer disclosures to AI uses that are not synthetic performers, such as AI-generated backgrounds, products, graphics or text, both to keep disclosures meaningful and to avoid confusing audiences. The goal is to catch genuine synthetic performers precisely, not to label everything. The fourth step is coordination with the wider disclosure ecosystem: the New York duty sits alongside platform AI-labelling policies and tool-level provenance metadata from regimes like California's, so rather than treating each as a separate task, advertisers should align them — a single identification of the AI performer can drive the New York disclosure, the platform AI label, and preservation of provenance together. Because advertisers usually have actual knowledge of how their creative was made, this production discipline directly addresses the law's trigger: knowing use of a synthetic performer means the disclosure is required, and a built-in checkpoint ensures it is never forgotten. Institutionalising the habit also avoids the repeat-violation penalty, which is higher than the first-violation penalty. For the US platform-compliance backdrop see the United States compliance guide , and screen creative and disclosures with the Keyword Risk Checker . The organizing principle is that advertisers comply by embedding a synthetic-performer checkpoint in production, disclosing conspicuously where one appears, not over-disclosing on other AI, and aligning with platform and provenance labels.

New York's Synthetic Performer Ad Disclosure Law 2026

Quick Answer

New York's synthetic performer disclosure law, codified at N.Y. General Business Law section 396-b, requires advertisers to conspicuously disclose when an advertisement contains an AI-generated synthetic performer, and its disclosure requirements took effect on June 9, 2026. The law applies to any person engaged in the business of dealing in property or services who creates or produces an advertisement for a commercial purpose, and the obligation is triggered where that person has actual knowledge that a synthetic performer appears in the advertisement. A 'synthetic performer' is defined as a digitally created asset — created, reproduced or modified by computer using generative AI or a software algorithm — that is intended to give the impression it is engaging in an audiovisual or visual performance of a human performer who is not recognizable as any identifiable natural person. In other words, it targets AI-generated fictional 'people' who appear to act or perform in an ad, not depictions of real, identifiable individuals. Failure to comply carries civil penalties of one thousand dollars for a first violation and five thousand dollars for subsequent violations. There are exceptions, including audio-only advertisements and cases where AI is used solely to translate the language of a human performer. Unlike provenance rules that bind AI tool providers, this is a direct duty on the advertiser producing the ad, which makes it a distinct compliance item: advertisers running visual or audiovisual ads to New York audiences must identify AI synthetic performers and add a conspicuous disclosure. Screen creative and disclosures with the Keyword Risk Checker, define terms in the compliance glossary, and track state AI rules on the Policy Change Tracker.

New York's Synthetic Performer Disclosure Law in 2026: What Advertisers Must Disclose About AI-Generated Performers

A First-in-the-Nation Advertiser Duty

New York's synthetic performer disclosure law, codified at N.Y. General Business Law section 396-b, is described as a first-in-the-nation measure requiring advertisers to disclose when an advertisement includes an AI-generated synthetic performer. Signed at the end of 2025, its disclosure requirements took effect on June 9, 2026. What makes it notable is that it places the obligation directly on advertisers — the businesses that create or produce ads — rather than on the AI tools used to make the content.

This distinguishes it from provenance-focused regimes that require AI systems to embed watermarks or offer detection tools. Those rules operate at the tool level; New York's law operates at the advertiser level, requiring a conspicuous disclosure in the ad itself when a synthetic performer appears. For any advertiser running visual or audiovisual campaigns that reach New York audiences and use AI-generated performers, it is a concrete, advertiser-facing compliance task with defined penalties, not a background technical standard.

"Requiring disclosure when advertisements include AI-generated synthetic performers gives consumers the information they need to know whether the person they are watching is real.
— Summary of New York's synthetic performer disclosure law"

This guide explains what qualifies as a synthetic performer, who must disclose and when, the exceptions and penalties, and how advertisers should build compliance into their creative process. Because it converges with other synthetic-media regimes, it should be read alongside the California AI Transparency Act guide and the EU AI Act Article 50 guide.

What Counts as a Synthetic Performer

The scope of the law turns on the definition of 'synthetic performer', which is precise and narrower than 'any AI-generated content'. Understanding exactly what falls within it is the key to knowing when the disclosure duty applies, because many uses of AI in advertising do not involve a synthetic performer at all.

The Definition's Elements

Element	Requirement	What it captures
Digitally created asset	Created, reproduced or modified by computer	Computer-generated or AI-modified assets
Made with AI or algorithm	Using generative AI or a software algorithm	AI or algorithmic generation, not filmed footage
Appears to perform	Gives the impression of an audiovisual or visual performance of a human	A 'person' who seems to act or perform on screen
Not a real, identifiable person	Not recognizable as any identifiable natural person	A fictional AI 'human', not a depiction of a real individual

The combined effect is that a synthetic performer is an AI-generated fictional human who appears to act or perform in an ad but is not a real, identifiable person. This deliberately excludes two adjacent categories: it does not target AI depictions of real, identifiable individuals — which raise separate right-of-publicity and likeness issues — and it does not target AI uses that do not create a performing 'human', such as AI-generated backgrounds, products, graphics or text. The performance-of-a-human element is central. For definitions of the underlying concepts, see the compliance glossary.

Who Must Disclose and When

The obligation is targeted by both who the actor is and what they know. The law applies to any person engaged in the business of dealing in property or services who creates or produces an advertisement for a commercial purpose, and the duty to disclose is triggered where that person has actual knowledge that a synthetic performer is in the advertisement.

The Two Conditions

Commercial advertiser producing the ad: the duty falls on a person in the business of dealing in property or services who creates or produces an advertisement for a commercial purpose — the advertiser or producer, not the platform that merely displays it.
Actual knowledge: the obligation attaches where the person has actual knowledge that a synthetic performer appears in the ad, tying the duty to what the advertiser actually knows about its own creative.
Conspicuous disclosure: where both conditions are met, the advertisement must conspicuously disclose that it contains a synthetic performer.

The 'actual knowledge' standard is significant: because advertisers typically know how their own creative was made, an advertiser that deliberately uses an AI synthetic performer will ordinarily have the knowledge that triggers the duty. The requirement that the disclosure be conspicuous means it must be noticeable to the audience, not hidden — consistent with disclosure standards elsewhere in advertising law. Because the law reaches advertisers producing ads for a commercial purpose, businesses running visual or audiovisual campaigns to New York audiences should assume it applies to their AI-performer creative. For the related endorsement-disclosure standards, see the FTC AI endorsement guide.

Exceptions and Penalties

The law contains specific exceptions that narrow its reach, and defined civil penalties that give it teeth. Knowing both helps advertisers judge when the duty does not apply and understand the cost of getting it wrong.

Exceptions and Consequences

Item	Detail
Audio-only exception	Audio-only advertisements are excluded from the requirement
Translation exception	Cases where AI is used solely for language translation of a human performer are excluded
First violation	Civil penalty of one thousand dollars
Subsequent violations	Civil penalty of five thousand dollars per violation

The exceptions reflect the law's focus on visual or audiovisual synthetic performers: an audio-only ad falls outside it, and using AI purely to translate a genuine human performer's speech into another language — where the performer is real and only the language is changed — does not create a synthetic performer requiring disclosure. The penalty structure, escalating from one thousand dollars for a first violation to five thousand dollars for subsequent ones, is modest per instance but can accumulate across campaigns and creates a clear incentive to build disclosure into the process rather than risk repeated violations. Track how comparable state measures develop on the Policy Change Tracker.

How Advertisers Should Comply

Compliance is straightforward once the duty is understood, because it reduces to a repeatable production step: identify whether any creative contains a synthetic performer, and if so, add a conspicuous disclosure. The challenge is operational consistency rather than legal complexity.

The Compliance Steps

Flag synthetic performers in production: build a check into the creative workflow that identifies when an ad uses an AI-generated performing 'human' who is not a real, identifiable person.
Add a conspicuous disclosure: where a synthetic performer appears in a visual or audiovisual ad, include a clear, noticeable disclosure that the ad contains a synthetic performer.
Distinguish from other AI uses: avoid over-disclosing on AI that is not a synthetic performer — backgrounds, products, graphics — while making sure genuine synthetic performers are caught.
Coordinate with platform and provenance labels: the New York disclosure sits alongside platform AI labels and tool-level provenance, so align them rather than treating each separately.

Because advertisers usually know how their creative was made, the actual-knowledge trigger means the practical safeguard is a deliberate production checkpoint: whenever AI generates a performing human figure, treat the disclosure as required for New York audiences. Building this into briefs and creative review — rather than catching it at launch — makes compliance routine and avoids the repeat violations that carry the higher penalty. The New York duty also dovetails with the broader synthetic-media disclosure trend, so a single habit of identifying and disclosing AI performers serves multiple regimes. For the US platform-compliance backdrop see the United States compliance guide, and screen disclosures with the Keyword Risk Checker.

Synthetic Performer Disclosure Checklist

[ ] Determined whether your visual or audiovisual ads reach New York audiences
[ ] Identified any creative using an AI-generated performing 'human' figure
[ ] Confirmed the figure is not a real, identifiable natural person
[ ] Added a conspicuous disclosure that the ad contains a synthetic performer
[ ] Confirmed the disclosure is noticeable, not hidden
[ ] Excluded audio-only ads and AI-translation-only cases where the exceptions apply
[ ] Avoided over-disclosing on AI uses that are not synthetic performers
[ ] Built a synthetic-performer check into the creative production workflow
[ ] Aligned the disclosure with platform AI labels and provenance metadata
[ ] Confirmed current section 396-b requirements against official New York sources

Don't miss the next policy change.

Create a free account — track every policy change across 8 platforms, get instant alerts, and access every free compliance tool. Or try our Meta Rejection Predictor first.

Create Free Account

Report Keywords — Run AI Compliance Audit

#AI Disclosure#New York#Synthetic Performer#Advertisers#Ad Compliance#Regulation#Generative AI#Deceptive Advertising#2026 Policy#Synthetic Media#Disclosure Rules#Compliance Guide 2026

Share This Report

Tweet Share

UK Online Safety Act Age-Assurance Enforcement 2026: How Ofcom's Ad-Network Business Disruption Lever Reaches Advertisers

Ofcom's Online Safety Act enforcement can reach beyond platforms to the advertising networks that fund them. With a July 2026 age-assurance effectiveness report due, here is how the business-disruption lever changes brand-safety calculus.

EU DSA Article 26 — Political Advertising Transparency: First-Year Implementation Data Across 27 Member States

EU DSA Article 26 governs political ad transparency across the EU — first-year data shows uneven member-state activity, Ireland enforcement concentration, and a tiered penalty structure.

FTC AI Endorsement Rules 2026 — 16 CFR Part 255 Application to Synthetic Content & State Equivalents

The FTC's 16 CFR Part 255 framework now applies to synthetic endorsements with per-violation penalties up to $53,088, stacking with California, Colorado, and New York laws into compound liability.

Platforms

Industries

Knowledge

Blog

New York's Synthetic Performer Disclosure Law in 2026: What Advertisers Must Disclose About AI-Generated Performers