On June 9, New York's synthetic performer disclosure law took effect. Any commercial advertisement using an AI-generated human likeness — a synthetic performer — now has to say so, conspicuously, where the creator knows it is being used. First violation runs $1,000, repeats $5,000. The law is narrow on its face: it is about disclosure, about telling the viewer that the face selling them something is not a real person. But narrow laws have a way of revealing wide problems, and this one exposes something performance teams have been able to ignore until now. The ad and the page have started making different promises, and nobody owns the gap between them.
What the news actually reveals
The synthetic performer law is the first regulation to treat AI ad creative as a thing that needs governance at the point of production. It puts liability on the party that creates the ad, not the platform that serves it. That is a meaningful shift. For two years, the story of AI in advertising has been pure velocity — Meta Advantage+, Google AI Max, TikTok Symphony, all racing to let teams produce more creative variants faster than any human could review them. The pitch was always volume, and the volume arrived.
The production side scaled. The governance side did not. Most teams now ship AI creative at a rate that no longer fits through a human approval queue, and the creative governance layer that used to be a brand manager eyeballing every asset has quietly disappeared. Nobody decided to remove it. It just stopped being able to keep up, and the velocity targets did not wait for it. The New York law does not care that you generated four hundred variants this month. It asks whether each one served to a New York consumer carries the disclosure. That is a per-asset compliance question applied to an asset base that grew past the point of per-asset human attention.
And disclosure is only the visible edge of the problem. Once an ad declares itself AI-generated, the landing page it points to has to honor that signal. A synthetic-presenter ad that drives to a page leaning on real human testimonials creates a tone-and-content contradiction — the ad admits it is synthetic, the page implies it is not. That is alignment drift with a compliance cost attached, and it is invisible to teams who check the ad and the page separately, which is nearly all of them. The disclosure rule did not create the inconsistency. It made one specific inconsistency expensive, which is usually how a latent problem becomes a budgeted one.
A cautious prediction
I would not bet the house on any single regulatory timeline, but the direction looks fairly clear over the next 6 to 12 months.
First, New York will not be alone. State-level AI advertising disclosure rules will likely follow the pattern of privacy law — a handful of states move, the compliance burden becomes national in practice because nobody builds separate creative for separate states, and a de facto standard emerges well ahead of any federal rule. Teams running national paid campaigns will probably end up treating disclosure as default rather than jurisdictional, because the cost of maintaining state-by-state creative variants is higher than the cost of disclosing everywhere. The path of least resistance points at a single standard.
Second, and more consequential for performance teams, the disclosure obligation will likely expose the post-click experience as the next governance frontier. Once you are forced to be honest in the ad about what is synthetic, the inconsistency between a disclosed-AI ad and a page that pretends otherwise becomes both a conversion problem and a legal one. The two stop being separable. Smart teams will probably respond by automating ad-to-page congruence checks the same way they automated creative production — because the only way to govern a system producing assets at machine speed is to measure it at machine speed. You cannot inspect a machine-speed pipeline with a human-speed process and expect the inspection to mean anything.
The mistake to avoid is the obvious one: scaling creative volume further without building the governance layer underneath it. Every team that responds to disclosure law by adding a single manual compliance reviewer is solving a machine-speed problem with a human-speed tool. That reviewer becomes the bottleneck, then gets overruled by the velocity targets, then the gap reopens — usually right before someone notices it in an audit. Hedge on the timeline if you like, but the structural pressure is not ambiguous. Production capacity and review capacity have diverged, and disclosure law is simply the first force loud enough to make the divergence show up on a balance sheet.
There is also a quieter risk in over-correcting. Some teams will read a law like this and pull back on AI creative entirely, treating the whole category as a liability to be minimized. That is probably an overreaction, and an expensive one — the production economics that made AI creative attractive have not changed, and competitors who keep producing at volume while governing it properly will pull ahead of the ones who retreat. The winning posture is not less creative; it is the same volume of creative with a measurement layer that can keep pace. Disclosure law rewards teams that can demonstrate control, not teams that produce less. Control at volume is the whole game now, and it is a measurement problem before it is a compliance one.
Where philosophy meets the spreadsheet
There is an old idea worth dragging into this. The unexamined campaign is not worth running. For most of the AI creative boom, teams have been operating on the hope that their ads and pages still agree with each other — hope, not measurement. It was cheaper to assume than to check, and nothing forced the issue. The synthetic performer law removes the comfort of that hope by attaching a dollar figure to one specific kind of disagreement, and once one kind has a price, the rest stop feeling free.
The deeper point is that knowing your ads and pages say the same thing is worth more than assuming they do, and it always was. Disclosure law just made the cost of the assumption legible. The teams that close the governance gap will not be the ones with the best AI generation tools — everyone has those now, and they are converging on the same capabilities anyway. The edge moves to whoever can prove, per asset, that the promise in the ad survives the click. That is what AdAlign measures, and it is why we think congruence stops being a CRO nicety and becomes the governance layer the AI creative stack forgot to build. The generation problem is solved. The agreement problem is wide open, and it is about to have regulators attached to it.
Where to start
If your creative volume is outpacing your ability to confirm that each ad still matches its page, you have a governance gap whether or not you can see it yet. Run a free audit to find out where your ads and pages disagree. Teams managing this across one brand should start monitoring at €99/mo; agencies governing it across many clients should look at the €299 Agency workspace. For the tactical version of this — how we fix message match in 48 hours — read our piece on AI Max and the landing page.
Frequently asked questions
What is New York's synthetic performer disclosure law? Effective June 9, 2026, it amends New York General Business Law to require advertisers to conspicuously disclose when a commercial advertisement uses an AI-generated synthetic performer — a digitally created human likeness — where the creator knows it is being used. Penalties are $1,000 for a first violation and $5,000 for repeats.
Who is liable under the synthetic performer law? Liability falls on the party that produces or creates the advertisement, not the platform that serves it. For most paid campaigns that means the brand and its agency, not Meta, Google, or TikTok.
How does AI ad disclosure relate to landing page alignment? Once an ad discloses that it is AI-generated, the landing page has to honor that signal. A synthetic-presenter ad pointing to a page that leans on real human testimonials creates a tone-and-content contradiction — alignment drift with a compliance cost, not just a conversion cost.
What is the creative governance gap? It is the gap between how fast teams now produce AI ad creative and their ability to confirm each asset still matches its landing page and meets disclosure rules. Production scaled to machine speed; the review layer stayed at human speed.
How do I keep AI ad creative compliant at volume? Automate the checks the same way you automated production: score every ad-to-page pair for congruence and disclosure consistency continuously, rather than relying on a manual reviewer who becomes a bottleneck and gets overruled by velocity targets.