That cheerful spokesperson enthusiastically praising a new skin cream, banking app, or miracle stain remover may look human, blink like a human, and deliver jokes with suspiciously perfect timing. In New York, however, advertisers can no longer assume audiences will simply accept that the person was born rather than rendered.
New York’s synthetic performer disclosure law took effect on June 9, 2026. Codified in Section 396-b of the New York General Business Law, it requires certain commercial advertisements to conspicuously disclose when they contain an artificial intelligence-generated “synthetic performer.” State officials have described it as a first-in-the-nation transparency measure for AI performers in advertising.
The law does not ban virtual actors, prohibit generative AI, or demand a warning label every time Photoshop has been opened. Instead, it focuses on a specific situation: a business knowingly uses a computer-generated human performer in an advertisement without making that performer recognizable as an actual person.
What Is New York’s Synthetic Performer Law?
The legislation, enacted as A.8887-B and S.8420-A, amended an existing New York advertising statute. Governor Kathy Hochul signed the measure in December 2025, and its disclosure requirements became effective 180 days later.
Under the law, a business dealing in property or services must conspicuously disclose the presence of a synthetic performer when it produces or creates a commercial advertisement and has actual knowledge that the performer appears in it. The requirement applies regardless of whether the advertisement runs on television, a website, a social media platform, a digital display, or another visual or audiovisual medium.
That wording matters. New York is regulating transparency rather than shutting down the technology. Brands may continue using AI-generated advertising talent, but they must tell viewers what they are seeing. Think of it as a name tag for the person who technically has neither a name nor a birth certificate.
What Qualifies as a Synthetic Performer?
New York defines a synthetic performer as a digitally created asset that is created, reproduced, or modified by computer using generative AI or another software algorithm. The asset must be intended to create the impression that it is giving a visual or audiovisual performance by a human who is not recognizable as an identifiable natural person.
Several elements must therefore be present:
- The asset was created or materially modified using a computer.
- Generative AI or a software algorithm was involved.
- The asset appears to be performing as a human.
- The apparent performer is not recognizable as a particular real person.
- The asset appears in a commercial advertisement.
Examples Likely to Require Disclosure
A fully generated spokesperson explaining the benefits of an insurance policy would probably fit the definition. So could a virtual fashion model walking through an AI-created showroom, a computer-generated chef demonstrating cookware, or an artificial customer enthusiastically describing a mattress that the “customer” has never slept on because the customer does not exist.
The definition can also reach assets that were reproduced or modified rather than created entirely from scratch. A campaign could therefore raise questions when a real performance is transformed so substantially that the finished person no longer appears to be an identifiable natural performer.
AI Uses That May Fall Outside the Law
The statute is not a universal labeling requirement for AI-assisted advertising. Generating a background, writing draft copy, removing image noise, resizing a banner, correcting colors, or creating an animated toaster with strong opinions about breakfast would not ordinarily create a synthetic human performer.
Nonhuman characters generally fall outside the central definition because they are not presented as human performers. Clearly fantastical creatures, talking animals, product mascots, abstract avatars, and animated objects are less likely to trigger this particular disclosure requirement.
The law also distinguishes synthetic performers from digital replicas of identifiable people. A generated person deliberately made to look like a particular celebrity, model, influencer, or customer may implicate publicity, privacy, contract, intellectual property, or consumer protection laws even when that person falls outside New York’s definition of a non-identifiable synthetic performer. In other words, escaping one definition does not mean the legal department gets the afternoon off.
Who Is Responsible for Making the Disclosure?
The obligation applies to a person or business that produces or creates the commercial advertisement and has actual knowledge that it includes a synthetic performer. Depending on the campaign structure, that group could include a brand, advertising agency, production company, creative studio, or another party exercising control over the ad’s creation.
The law does not rely on a broad “should have known” standard. It refers to actual knowledge. Nevertheless, advertisers should not treat that language as permission to avoid asking vendors how creative assets were made. Deliberately maintaining a mysterious black box marked “probably not robots” is not a durable compliance strategy.
Contracts with AI vendors, production studios, influencers, agencies, and freelancers should require disclosure of synthetic content. Approval systems should also record who created each asset, which tools were used, whether human source material was involved, and who determined that a disclosure was or was not necessary.
Media Platforms and Publishers Are Treated Differently
The statute generally excludes advertising media that merely publish or distribute a noncompliant ad. Its language identifies newspapers, magazines, television networks, stations, streaming services, cable systems, billboards, and transit advertising among the covered distribution channels that are not responsible solely because they disseminated the material.
That exemption does not necessarily protect a media company that helps create the advertisement. A streaming service acting only as a distribution platform occupies a different position from one that develops a synthetic spokesperson as part of its own branded campaign.
What Does “Conspicuously Disclose” Mean?
New York’s law requires a conspicuous disclosure but does not prescribe exact wording, font size, screen position, color, duration, or technical format. It does not hand advertisers a magic compliance sticker that can be pasted into the lower-right corner at three pixels tall.
Until courts or regulators provide more detailed guidance, advertisers can look to established Federal Trade Commission principles for clear and conspicuous disclosures. FTC guidance emphasizes that disclosures should be placed close to the relevant content, presented prominently, remain visible long enough to be noticed, and use language ordinary consumers can understand. The disclosure should also work on every device and platform where the advertisement appears.
Practical Disclosure Language
Possible wording could include:
- “This advertisement contains an AI-generated synthetic performer.”
- “The person shown in this advertisement was created using artificial intelligence.”
- “AI-generated performer: This is not a real individual.”
Video disclosures should be large enough to read, remain visible for a meaningful period, and contrast with the background. A label that disappears faster than a limited-time coupon may not accomplish much.
For social media content, the notice should appear in the creative itself rather than being buried after twenty hashtags, hidden in a collapsed caption, or placed on a separate landing page. For static advertisements, the disclosure should be reasonably close to the synthetic person. Requiring viewers to locate a microscopic footnote is not transparency; it is an eyesight test.
Penalties for Violating the New York AI Advertising Law
A violation can produce a civil penalty of $1,000 for a first offense and $5,000 for each subsequent offense. The law does not expressly create a private right of action under this particular disclosure provision, so enforcement is expected to come primarily from New York state authorities.
The statute does not fully explain how separate violations will be counted. An unanswered question is whether one creative asset distributed through numerous placements will be treated as one violation, several campaign violations, or multiple violations based on individual advertisements. That uncertainty gives businesses another reason to address disclosure during production rather than after a campaign has accumulated millions of impressions.
The financial penalty may also be only one part of the risk. A brand accused of hiding AI-generated talent can face campaign delays, replacement costs, contractual disputes, public criticism, and reduced consumer trust. Re-editing hundreds of localized videos is rarely the sort of surprise that makes a marketing director’s Monday more enjoyable.
Important Exemptions and Limitations
Advertising for Expressive Works
Advertisements and promotional materials for expressive works can be exempt when the synthetic performer’s use in the promotional material is consistent with its use in the work itself. Expressive works named in the statute include motion pictures, television programs, streaming content, documentaries, video games, and similar audiovisual productions.
For example, if a synthetic character appears in a science-fiction film, a trailer showing that same character consistently with the film may fall within the exemption. Creating an unrelated synthetic salesperson solely to promote the film presents a different question.
Audio-Only Advertisements
Audio-only advertising is expressly excluded. The law’s definition and disclosure system are directed primarily at visual and audiovisual human performances. Separate laws may still apply to cloned voices, deceptive testimonials, impersonation, or unauthorized use of an identifiable person’s voice.
AI Used Solely for Translation
The statute also excludes situations in which AI is used solely to translate the language of a human performer. The exemption is especially relevant to campaigns localized for multiple markets. It may not apply when the process goes beyond translation and generates a materially different performance or artificial spokesperson.
Ordinary Distribution Channels
Media outlets and platforms are generally exempt when their role is limited to publishing or disseminating an advertisement created by another party. The statute also states that it does not expand or reduce protections provided to interactive computer services under Section 230 of the federal Communications Decency Act.
Why New York Adopted the AI Performer Rule
Supporters argue that consumers deserve to know whether a persuasive human presence in an advertisement is genuine. Synthetic presenters can imitate eye contact, emotional expression, conversational pauses, and personal enthusiasm. Those qualities can create an implied sense of experience or credibility even though the “performer” has never bought, tasted, worn, driven, or tested anything.
SAG-AFTRA supported the legislation as a safeguard for transparency and the economic value of human performance. The union described the law as a narrowly tailored response to synthetic humans used in advertising. Advertising industry groups, including the 4As, raised concerns that broad definitions and uncertain disclosure standards could complicate production and discourage innovation.
Both perspectives reflect genuine concerns. Performers do not want their work silently displaced by endless digital talent, while advertisers need rules that distinguish misleading synthetic people from harmless creative tools. New York’s law attempts to strike that balance by permitting the technology while requiring viewers to be informed.
How the Law Interacts With Other AI and Performer Protections
The synthetic performer disclosure law is only one piece of New York’s developing framework. Other state protections address digital replicas of models, unauthorized commercial use of identifiable people, and digital representations of deceased performers.
New York’s Fashion Workers Act, for example, requires separate written consent for certain uses of a model’s digital replica. Such consent should address the purpose, scope, duration, and compensation associated with the replica. New York has also expanded protections involving unauthorized digital representations of deceased personalities and performers.
Federal advertising rules remain relevant as well. The FTC has explained that AI-generated avatars are not automatically prohibited, but their use can become deceptive when they deliver false testimonials, misrepresent consumer experience, or make viewers believe that an identifiable celebrity endorsed a product without permission.
A compliant New York disclosure therefore does not cure every possible legal problem. Labeling an avatar “AI-generated” will not make a false health claim truthful, transform a fabricated customer experience into a genuine testimonial, or authorize the use of someone else’s likeness.
Compliance Steps for Advertisers and Agencies
- Inventory AI-assisted advertising. Identify campaigns containing generated or materially modified humans.
- Classify each asset. Determine whether it presents a human performance, a real identifiable person, a nonhuman character, or ordinary edited footage.
- Document actual knowledge. Record vendor disclosures, production methods, source files, approvals, and internal legal reviews.
- Add the notice inside the ad. Avoid relying exclusively on metadata, external webpages, platform descriptions, or hidden captions.
- Test every format. Confirm that the disclosure remains readable in vertical videos, mobile placements, banners, television spots, and shortened edits.
- Update contracts. Allocate responsibility among brands, agencies, production companies, and technology vendors.
- Review related laws. Check testimonial rules, publicity rights, model consent, privacy, copyright, and state-specific AI requirements.
Practical Experiences: What Advertising Teams Learn During AI Performer Reviews
Experience One: The Technology Team and Legal Team Use Different Dictionaries
A recurring campaign experience begins with someone saying, “It is not really AI-generated; we only used AI to modify it.” That sentence may be technically meaningful to a developer, but New York’s definition includes assets that are created, reproduced, or modified by computer. The practical lesson is to ask what changed in the final human performance rather than arguing about the label attached to the software.
A structured intake form helps enormously. It should ask whether a real person was recorded, whether facial movement was generated, whether the voice came from a human, whether the performer resembles a known individual, and whether the finished asset could reasonably be perceived as a real human performance.
Experience Two: One Master Video Quickly Becomes Fifty Versions
A campaign may begin with one compliant video containing a readable disclosure. The marketing team then crops it for vertical social media, cuts it to six seconds, adds subtitles, removes the opening frame, and converts it into a banner. Somewhere during that enthusiastic optimization festival, the notice vanishes.
The safest workflow treats the disclosure as a mandatory creative element, much like a logo or legally required product statement. It should be included in templates and checked again after every resize, localization, edit, and platform conversion.
Experience Three: Vendor Transparency Is More Valuable Than Vendor Confidence
Creative vendors may confidently promise that their avatars are “fully compliant,” but compliance depends on where the advertisement runs, how the performer was generated, and how the notice appears. A contractual warranty is useful, yet it cannot replace technical information.
Experienced teams request an asset-development record describing the tools used, the source materials, any human training data supplied by the advertiser, and whether the avatar resembles an identifiable person. This record allows the brand to make its own decision and provides evidence of a serious compliance process.
Experience Four: Clear Language Usually Beats Clever Language
Creative teams sometimes want disclosures such as “Imagined with tomorrow’s technology” or “Enhanced by digital innovation.” Those phrases sound elegant and reveal almost nothing. A consumer may interpret them as references to visual effects, editing software, or an ambitious intern.
Direct wording performs better: “This ad contains an AI-generated performer.” The sentence may not win a copywriting award, but it answers the relevant question immediately. Compliance language has one job, and it is not to audition for a poetry festival.
Experience Five: Disclosure Review Often Reveals Bigger Problems
While reviewing an AI performer notice, teams frequently uncover unrelated risks. The avatar may repeat an unverified product claim, deliver a fabricated testimonial, imitate a celebrity’s speaking style, use copyrighted material in the background, or appear to represent a licensed professional.
This is why mature review programs do not treat the New York rule as a single checkbox. They evaluate the complete impression created by the advertisement. Consumers may understand that a performer is synthetic while still being misled about whether the performer’s statements reflect real experience, scientific evidence, or professional expertise.
The broader experience is encouraging: transparency rarely ruins a good campaign. A straightforward notice can demonstrate that a brand is using new technology responsibly. Consumers are becoming familiar with AI-assisted media, and many will not object to synthetic talent when the presentation is honest. The real reputational danger often comes from appearing to hide the technology rather than from using it.
Conclusion
New York’s AI performer advertising law marks a shift from voluntary disclosure to enforceable transparency. Beginning June 9, 2026, businesses that knowingly create commercial advertisements containing covered synthetic human performers must make that use conspicuous to viewers.
The rule is narrower than its headline may suggest. It does not regulate every AI-generated image, prohibit virtual talent, or automatically cover ordinary editing. It focuses on computer-created or materially modified assets that appear to deliver human performances without representing an identifiable natural person.
Advertisers should build disclosure decisions into creative development, vendor management, contracting, localization, and final quality assurance. Waiting until an advertisement is already running can turn a simple sentence into an expensive campaign repair project.