TL;DR: New York now requires a conspicuous disclosure when an ad shown to its audiences features an AI-generated human, called a synthetic performer. The law took effect June 9, 2026, with civil penalties of $1,000 to $5,000 per violation (New York Governor's Office, 2026). It doesn't ban AI in ads. It regulates honesty about people who aren't real.
On June 9, 2026, New York became the first state in the country to require advertisers to label AI-generated humans in their ads. Not AI in general. Not retouching, not set extensions, not generated backgrounds. People. If a person in your ad was created by a machine and a New Yorker might see that ad, the ad now has to say so.
I run a hybrid production studio in New York. We shoot real food, real drinks, and real people, and we use AI to extend that work into more formats and more campaigns. So this law sits directly on top of my desk. Here's what it actually says, where the lines are, and what brands and agencies should do about it, written from the production side rather than the legal side.
What Does New York's AI Ad Disclosure Law Require?
The law requires any ad featuring a synthetic performer to "conspicuously disclose" that fact when the ad reaches New York audiences. It was signed by Governor Kathy Hochul on December 11, 2025 and took effect June 9, 2026 (New York Governor's Office, 2025). New York describes it as the first law of its kind in the nation.
The mechanics are simple. The legislation, passed as S.8420-A in the Senate and A.8887-B in the Assembly, amends New York's General Business Law. The Senate passed it 59 to 0, which tells you something about where transparency rules are heading politically.
Two details matter for anyone running campaigns. First, the obligation applies when the advertiser has actual knowledge that a synthetic performer appears in the ad. Second, the statute doesn't define what "conspicuous" means. Advertising counsel are pointing clients to the FTC's clear-and-conspicuous standard as the safe benchmark: visible, readable, on screen long enough to register, and not buried in fine print (McDermott, 2026).
To be clear, the law doesn't ban anything. You can run a fully AI-generated spokesperson in Times Square if you want. You just have to tell people she isn't real.
What Counts as a Synthetic Performer?
A synthetic performer is a digitally created asset, made or modified using generative AI or an algorithm, that creates the impression of a human performance by someone who isn't a recognizable real person (S.8420-A bill text, 2025). In plain terms: an AI-generated human who looks like they're performing in your ad.
That definition is narrower than "AI-generated content" and wider than "fake celebrity." A few practical readings:
- A fully generated AI model holding your product is a synthetic performer. Disclose.
- A generated background extra walking past your storefront likely counts too. Legal analysts note the law covers synthetic performers appearing in any capacity, including background roles, and some companies are conservatively disclosing even AI extras (Reed Smith, 2026).
- A real model you photographed, with an AI-extended set behind her, is not a synthetic performer. The human performance is real.
- AI-generated food, drinks, products, and environments with no humans in them sit outside this law entirely.
Notice what the statute is actually protecting. It's not policing tools. It's policing the impression that a human being performed when no human did. That's a much cleaner line than most AI policy conversations manage, and it's a workable one for production teams.
Who Needs to Pay Attention?
Any brand, agency, or production company whose ads could reach New York audiences. That's effectively every national campaign and most regional ones. The law covers ads in any medium except audio-only formats, which puts digital, social, streaming, broadcast, print, and out-of-home all in scope (Reed Smith, 2026).
The timing isn't accidental. AI-generated creative has moved from experiment to default fast. The IAB found 86% of ad buyers are using or planning to use generative AI to build video ad creative, and roughly 30% of ad creative is already built or enhanced with generative AI, projected to approach 40% in 2026 (IAB, 2025). Regulation tends to show up right when a practice stops being niche.
The Out-of-State Trap: No Geofence, No Exemption
Here's the scenario that catches brands off guard. You're based in Texas. You run an Instagram campaign with an AI-generated spokesperson. You never think about New York, and you don't geofence your delivery. That ad is now distributed in New York State, and the disclosure requirement applies to it.
The law is triggered by where the ad is displayed or distributed, not where the advertiser is headquartered (Reed Smith, 2026). A national social campaign without geographic restrictions includes New York by default. That leaves three honest options: add the disclosure, geofence New York out of delivery, or use real performers. For most brands, designing the disclosure once is cheaper than fragmenting the media plan.
The "actual knowledge" standard deserves attention from agency producers specifically. If you're assembling campaigns from vendor-supplied assets, stock libraries, and freelance contributors, what you know about each asset's origin becomes a legal question. Counsel are recommending vendor warranties and asset inventories for exactly this reason. The brands in the best position are the ones who can answer, for any frame in any ad, what was photographed and what was generated.
That's not a legal burden so much as a production discipline. Studios that already track usage rights and asset provenance, the same way we track usage rights on commercial shoots, can answer that question in minutes.
What Are the Penalties?
Civil penalties run $1,000 for a first violation and $5,000 for each subsequent violation, with no private right of action (S.8420-A, 2026). For a national brand, those numbers are small. The real exposure isn't the fine. It's getting publicly caught passing off a synthetic person as real after the law made the standard explicit.
Here's the part most coverage misses: disclosure isn't just downside protection. A Yahoo and Publicis Media survey found that when consumers noticed an AI disclosure in an ad, ad trustworthiness rose 73% and overall trust in the company nearly doubled, up 96% (Yahoo, 2024). The same research found 72% of consumers say AI makes it hard to tell what's authentic.
Put those two findings together. Consumers assume AI is everywhere and can't tell where. Brands that label it honestly get a trust premium. New York just made the honest position mandatory, which means early, well-designed disclosure is now both compliance and positioning. The brands treating this as a creative problem, designing disclosures that fit the brand voice instead of slapping on legal boilerplate, will come out ahead of the ones treating it as a fire drill.
There's a gap worth knowing about here too. Advertisers and audiences see this differently: 77% of advertisers view AI in ads positively, against only 38% of consumers (IAB, 2025). If your team's instinct says nobody cares about disclosure, that instinct is measuring the wrong room.
Are There Exceptions?
Yes, four meaningful ones. The law carves out ads for expressive works such as film, TV, streaming content, and video games, where the synthetic performer's use matches the underlying work. It also exempts audio-only ads, AI used solely to translate a human performer's speech into another language, and publishers or platforms that merely distribute ads they didn't create (Cooley, 2026).
The expressive-works carve-out is the practical one for entertainment marketing. A trailer for a film that features digital characters doesn't need a synthetic performer label, because the ad reflects the work itself. The translation exemption matters for global campaigns: dubbing a real spokesperson into Spanish with AI doesn't trigger disclosure, because the human performance is real.
Platforms get a cure window. A distributor that's notified it's running a noncompliant ad has five days to act before facing liability (Reed Smith, 2026). The compliance weight sits with whoever made and placed the ad, not whoever displayed it.
One adjacent law worth flagging: New York signed a companion bill the same day requiring consent from heirs or executors before commercial use of a deceased person's digital replica (Skadden, 2026). If your creative plans ever touch archival talent or legacy brand ambassadors, that one belongs on your radar too.
Is New York the Only Law Like This?
No, and that's the strategic point. New York moved first on ad-specific disclosure, but two bigger frameworks arrive within weeks. The EU AI Act's Article 50 transparency obligations, which require labeling AI-generated and manipulated content, apply from August 2, 2026, with fines up to 15 million euros or 3% of global turnover (EU AI Act, 2026). California's AI Transparency Act takes effect the same day.
The FTC has been explicit that federal advertising law already applies regardless of how content was made. "There is no AI exemption from the laws on the books," as the agency put it when it launched its AI enforcement sweep (FTC, 2024).
So the question facing marketing teams isn't whether to build a disclosure process for New York. It's whether to build one process that satisfies New York in June, then Brussels and Sacramento in August, then whatever comes after. Brands running campaigns across NYC, London, and the EU should treat summer 2026 as the moment disclosure became a standing operational requirement, not a one-state quirk.
I covered the broader regulatory picture, including the UK ASA's position and the FTC's net impression standard, in my piece on the ethics of AI in food photography. The short version holds: every framework converges on the same question. Does the image create an expectation reality can't meet?
What Does This Mean for Hybrid Production?
Here's the reading that matters if you produce visual content: the law draws its line exactly where disciplined hybrid production already operates. AI-generated humans trigger disclosure. AI-extended photography of real humans, real food, and real places doesn't, because nothing synthetic is performing.
That distinction has been the architecture of my AI Studio work from the start. We photograph the real subject, with real light and real texture, then use AI to extend frames, build format variations, and multiply one shoot into a season of assets. The performance, when there is one, belongs to a person who was actually on set. Under New York's law, that work carries no disclosure obligation at all.
It also means your shoot documentation just became a compliance asset. Because the law turns on actual knowledge of what's synthetic, a production partner who can tell you precisely which assets were photographed, which were extended, and which were generated gives your legal team something most AI pipelines can't: a clean answer. We already maintain that record for licensing reasons. Now it does double duty.
None of this is an argument against synthetic performers. For some campaigns they'll be the right call, and the disclosure will be a fair trade. It's an argument for knowing which tool you're using and why, before the media plan locks.
How Should Brands Respond?
Start with an audit, not a policy memo. Most teams don't actually know how much generated humanity is sitting in their active creative. The practical sequence:
- Inventory active and upcoming campaigns that could reach New York. Flag any asset containing an AI-generated or AI-altered human figure, including background figures and partial performers.
- Ask your vendors directly. Stock libraries, production partners, and freelancers should be able to state what's generated. Get it in writing. The actual-knowledge standard makes vendor warranties your first line of protection.
- Design the disclosure once. Work out placement, wording, and duration that satisfy a clear-and-conspicuous reading and still respect the creative. Then template it across formats.
- Build the check into approvals. A single line item at final review, "any synthetic performers, and is the disclosure present," costs nothing. Correcting distributed media costs a lot more.
- Decide where you stand creatively. If real people and real product are central to your brand promise, say so. The trust data suggests audiences reward it.
If you're weighing how a campaign should split between shooting and generating, that's a conversation worth having before creative development starts, not at legal review. Book a call and we'll walk through how we structure hybrid campaigns that scale without crossing the lines this law just drew.
The disclosure era of AI advertising started June 9. The brands that treat transparency as part of the craft, rather than a constraint on it, are going to find this a very easy law to live with.
