New York has become the first US state to require advertisers to disclose AI-generated people in commercial ads. The law takes effect 9 June 2026 and applies to any ad shown to New York consumers, regardless of where the brand or agency sits. Penalties are $1,000 first violation and $5,000 each subsequent. The fix is a four-word label, but it requires a cross-channel inventory and a single standardised disclosure treatment before the deadline.

If you are a DTC brand running paid media into the US in 2026, there is a new compliance line you need to cross before June 9. New York has become the first US state to require advertisers to disclose when an ad uses an AI-generated person, and the law applies to any ad shown to New York consumers regardless of where the brand or agency sits.

AI-generated people now sit in almost every $5M+ DTC ad library, and the cross-channel audit most brands need to run before the deadline has barely started. We mapped this for our partners the week the statute was signed in December. This is the version of the compliance playbook we are walking founders through right now: what the law actually says, what counts and what does not, who is on the hook for enforcement, and the five steps to take before the deadline hits.

(For the broader Q2 platform context, Meta's Q2 2026 targeting changes is the companion piece on what is shifting across the rest of the channel.)

What the law actually says

On 11 December 2025, New York Governor Hochul signed S.8420-A / A.8887-B, an amendment to New York's general business law that requires advertisers to disclose the use of "synthetic performers" in commercial advertisements. It is the first law of its kind in the United States, and New York has openly positioned it as a template other states (California most obviously) will copy.

The statute defines a synthetic performer as a "digital asset created or modified by a computer using generative AI or a software algorithm that's meant to give the impression it's a real human in a visual or audiovisual performance." In plain terms, this is any AI-generated person in your ad creative. It covers both images and video.

The disclosure must be "conspicuous." The statute does not dictate exact wording, size, or placement, which sounds like flexibility but in practice is not. Enforcement will almost certainly be judged against the standards already in place for paid influencer disclosures, which require the label to be prominent and unavoidable to a normal viewer. A small grey watermark camouflaged in the corner is asking for a fine.

Penalties are $1,000 for a first violation and $5,000 for each subsequent one. The rule does not apply to audio-only ads, to AI used solely for language translation, or to ads for expressive works such as films, TV, and games where the synthetic performer also appears in the underlying work.

What counts and what does not

This is where the practical line matters, because not every AI-touched asset is in scope.

The rule is about AI-generated humans. If your ad shows an AI-generated person (a model holding your product, a synthetic spokesperson, a generated lifestyle shot of a "customer"), in either an image or a video, it needs a disclosure. If your ad is a product-only shot, an AI-generated background, or a wide environmental scene with no AI-generated person visible, it sits outside the requirement.

This is a meaningful distinction. AI models are still entirely usable in DTC ad creative. The law does not ban anything. It just asks brands to be honest about which assets are AI-generated. AI products and AI backgrounds are unaffected. AI people require a label.

That single rule, taught cleanly to your creative team this week, will cover roughly 95 percent of the practical compliance work.

Why this is bigger than Meta

A lot of regulatory thinking in DTC defaults to "what does Meta want us to do" because that is where most spend lives. This law cuts across that mental model. It applies to the advertisement in any media in which the advertisement appears. The same AI-model creative running across Meta, Google, YouTube, TikTok, Amazon, programmatic display, and connected TV needs the disclosure on every placement that reaches New York consumers, not just inside the Meta Ads Manager.

The practical implication is that compliance is a cross-channel audit, not a single-platform fix. Brands that handle this cleanly will be the ones running a single disclosure standard applied consistently across every placement. Brands that try to fix it channel by channel will end up with inconsistent labelling that varies by surface, which is a faster path to a fine than no labelling at all.

Who is on the hook

Three points worth being clear on, because the questions land in this exact order whenever we walk founders through this.

It applies regardless of where you are based. The law reaches any ad shown to people in New York, regardless of where the brand or agency sits. A brand headquartered in London or Austin running national US campaigns through Meta is in scope the moment a New York resident sees the ad. Geo-fencing New York out of your campaigns is a technical option, but for most $5M to $30M DTC brands targeting the US, the more practical answer is to comply across the whole US footprint and treat the disclosure as a standard label.

The platforms are not the enforcer. The compliance obligation sits on whoever produces or creates the advertisement, which means the brand and its agency. Meta is not going to build a magic disclosure toggle and Google is not going to police this on your behalf. Enforcement runs through civil penalties at the state level. The platforms will keep serving AI-generated creative without flagging it, because the regulatory responsibility is not theirs.

It is not actually that grey. On the format of the disclosure, yes, the statute leaves room for interpretation. On whether you need one at all for an AI-generated human, it is not grey. You do.

What to do before June 9, 2026

Five concrete moves, in order.

1. Inventory every live ad creative across every paid channel. Pull every active and queued ad across Meta, Google, YouTube, TikTok, Pinterest, Amazon, programmatic, and CTV. Flag anything containing an AI-generated person in an image or video. Pause flagged ads while you decide next steps. The audit itself usually takes a day. The political work of getting the agency to do it properly takes longer.

2. Decide relabel or replace, asset by asset. For each flagged creative, choose to add a clear AI disclosure or swap the asset for real talent or UGC. This is also a clean moment to question whether the AI model was actually earning its keep. If the asset was a placeholder rather than a winner, the compliance deadline is a useful prompt to retire it.

3. Standardise on one disclosure treatment. Build a single, legible label and use it consistently. Most brands we are working with are landing on a short label such as "AI-generated" in a bottom corner overlay, with high enough contrast to remain readable across every placement and every aspect ratio (1:1, 4:5, 9:16). The exact wording is less important than the consistency. Per-designer guesswork is what produces enforcement risk.

4. Update agency, production, and vendor agreements. Spell out who is responsible for applying the disclosure, and confirm whether any creative partner is quietly using synthetic performers in assets they hand you. Add indemnities where it makes sense. For brands using multiple production partners (in-house team, external agency, freelance designers, UGC platforms), this is the most under-managed compliance gap, and it is the one most likely to produce a missed asset.

5. Add an AI-human checkpoint to your pre-launch QA. Update your brand guidelines and review process so nothing ships into New York exposure without being cleared. Make sure the wider team understands the AI-human-versus-product distinction. A 60-second checkpoint on the QA template is the difference between a system that catches the issue and one that depends on individual vigilance.

If you do these five things before 9 June, you will be ahead of the vast majority of DTC brands on this compliance line. Most are still in step one.

What the disclaimer actually looks like in practice

The statute leaves the format open, which is why it helps to see what a working disclaimer actually looks like in a live creative. Three things matter: the wording, the placement, and the surface it sits on.

Wording

The version we use across client accounts is short, clear, and unambiguous about what is and is not real:

"Some visuals or voices are AI-generated. No real people or endorsements."

That sentence covers the synthetic performer requirement, makes the absence of real endorsements explicit (useful in any creative that could be confused for influencer content or UGC), and reads as a normal disclaimer rather than a warning. The exact wording can flex, but the two elements (AI-generated, no real people) should both appear.

Placement: in-creative or primary text copy

Two practical options.

The in-creative option means baking the disclaimer into the visual asset itself, typically as a text overlay in a bottom band. This is the most defensible interpretation of "conspicuous" because the disclosure is unavoidable for any viewer of the ad. It also travels cleanly across channels, which matters when the same creative is running on Meta, TikTok, YouTube, programmatic, and CTV simultaneously.

The primary-text-copy option means placing the disclosure in the body copy of the ad rather than on the creative itself. This is the lighter-touch approach and we have seen strict legal teams green-light it in client accounts. It is arguably less robust if a paid social asset is later repurposed onto a channel that does not carry primary text (CTV, programmatic display), so we lean toward in-creative as the default unless a brand has a clear reason to keep the creative clean.

Surface and legibility

The disclaimer needs to remain legible across every aspect ratio (1:1, 4:5, 9:16) and on small screens. White text on a dark band, or dark text on a cream band, with a font size at least matching the rest of any on-creative copy. Most failures we see are not the wording. They are the disclaimer rendered so small at 9:16 that it stops being conspicuous in any practical sense.

Performance impact (the part founders quietly ask about)

The honest answer, based on what we have been discussing with other founders and operators since the statute was signed, is that a properly placed AI disclaimer does not materially hurt creative performance. The pattern reads identically to what brands already see with FDA health-claim disclaimers and other compliance notices: viewers ignore the boilerplate, the algorithm does not penalise it, and the creative carries its own weight. There is no real performance trade-off here, only a compliance one. The operators with the longest compliance muscle (health, supplements, telehealth) have been treating this kind of disclosure as standard for months without seeing a hit, which is the cleanest real-world evidence available on the question.

A note on enforcement realism

Worth saying clearly so the panic-to-action ratio stays calibrated. The statute does not give individual consumers a private right of action. Enforcement runs through the New York Attorney General only, with a $1,000 first violation and $5,000 per subsequent violation. That removes the class-action and predatory-litigation vectors that have made ADA compliance and Prop 65 such painful spaces for DTC brands. The probability of being randomly targeted out of the gate is meaningfully lower than the headline penalty suggests. None of this is a reason to skip the work. It is a reason to do the work without panic.

The bigger picture: what comes next

New York is first, not last. California has been the obvious next state since the statute was signed, and there are early signals that EU and UK regulators are watching this space closely as part of broader AI labelling discussions.

For founder-led DTC brands at $5M to $30M, this is the moment to build the operational habit rather than treat it as a one-off compliance task. Brands that run a single AI disclosure standard across all channels, with vendor agreements that spell out responsibility, and with a QA checkpoint that catches issues before launch, will not be fazed when the next five states pass similar laws. Brands that scramble on a state-by-state basis will spend the back half of 2026 in reactive mode while their competitors compound.

It is also worth saying clearly: this law is not a reason to stop using AI creative. The productivity engine is real. Brands using AI-generated people in ads are moving faster, testing more, and producing visual variety that brands relying on traditional shoots cannot match. (We covered the algorithmic side of why creative variety matters in the Andromeda Q2 update and the First $100k of Meta Spend playbook.) The law just asks you to be transparent about which assets are AI-generated. That is a four-word label in the corner of a frame, not a strategic pivot.

Our honest take

Transparency about AI usage is becoming a competitive advantage rather than a regulatory burden. In a feed where AI-generated content is increasingly common, the brands treating "this is AI" as a normal part of the asset, rather than something to hide, will earn more trust than the ones being caught out by enforcement.

The brands that compound from here will be the ones who treat compliance as an operational discipline, not as a panicked retrofit. Five steps. Three weeks. Do them now and you will have built a system that absorbs the next set of state-level rules with no additional scramble.

Where to go next

Webtopia runs paid media for DTC brands at $5M to $30M across Meta, Google, Pinterest, and programmatic. Our creative process now includes the AI disclosure standard as part of pre-launch QA on every asset. If you want a view on whether your current ad library is exposed before 9 June, book a call and we will walk through your creative inventory with you.

For the broader Q2 2026 platform context, Meta's Q2 2026 targeting changes is the companion piece. For the reporting context that determines how you measure the AI creative that is performing for you, MER vs ROAS in 2026 is the long-form piece.

Frequently asked questions

What is a synthetic performer under New York's AI advertising law?

A synthetic performer is defined in New York's S.8420-A / A.8887-B as a digital asset created or modified using generative AI or a software algorithm, intended to give the impression of being a real human in a visual or audiovisual performance. In plain terms, it is any AI-generated person appearing in ad creative, in either an image or a video. AI-generated products, backgrounds, and environmental scenes without AI-generated humans are not synthetic performers and do not require disclosure.

When does the New York AI ad disclosure law take effect?

The law takes effect on 9 June 2026. Brands and agencies running paid media targeting New York consumers need to have their AI disclosure compliance in place before that date. Penalties are $1,000 for a first violation and $5,000 for each subsequent violation.

Does the law apply to my brand if I am not based in New York?

Yes. The law reaches any ad shown to people in New York, regardless of where the brand or agency is based. A brand headquartered in London, Austin, or anywhere else running US-targeted campaigns through Meta, Google, YouTube, TikTok, Amazon, or programmatic display is in scope the moment a New York resident sees the ad. Compliance is required across every channel where the ad appears.

Do I need to disclose AI-generated backgrounds or product images?

No. The disclosure requirement applies only to AI-generated humans (synthetic performers) in ad creative. AI-generated product shots, AI-generated backgrounds, and AI-generated environmental or wide scenes without human figures do not require the disclosure. The line is whether an AI-generated person is visible in the asset.

Who is responsible for compliance, the platform or the advertiser?

The advertiser. The compliance obligation sits on whoever produces or creates the advertisement, which means the brand and its agency. Meta, Google, TikTok, and other platforms are not the enforcers and will not police disclosure on your behalf. Enforcement runs through civil penalties at the state level. Brands should update their agency and vendor agreements to spell out who applies the disclosure and where the responsibility sits.