New York State has banned AI influencers in ads. 

Here’s everything you need to know.

On December 11, 2025, Governor Kathy Hochul signed S.8420-A/A.8887-B, which requires anyone producing or creating advertisements that include “AI-generated synthetic performers” to conspicuously disclose that fact. This is the first ever law like this in the USA. 

The law takes effect on June 9th. Gird your loins. 

The law applies to every paid advertising channel an ad runs on (Meta, Google, YouTube, TikTok, CTV, display, anything) if the ad is shown in New York state. It doesn’t matter where YOU are located: it matters if the ad is received inside New York state borders. 

The trigger is narrow. Disclosure is required only when an ad uses a "synthetic performer": a digitally created asset made or modified using generative AI or a software algorithm that's intended to create the impression of a human performer who isn't recognizable as any identifiable natural person. 

I think under this interpretation, you could make AI versions of existing, named humans. Just not unnamed, not-real persons from scratch. 

Therefore this law bans ads featuring influencers like Aitana Lopez, the AI-generated supermodel with hundreds of thousands of followers, shown in the header image of this post. 

What “disclosure” has to look like is unclear. I’m assuming tiny text on screen, like we see for pharmaceutical ads on television. This is recommended by most law blogs I’m reading. 

The law doesn’t apply to all AI content. It does not regulate AI in advertising generally or require disclosure for all AI-generated imagery.

AI-generated product shots or backgrounds without humans don't need the label, and it doesn't cover cartoon characters, animals, or fantastical figures not meant to represent a human. So maybe those brainrot Pixar-style ads are the future? Please no. Please god, no. Noooo. 

There are exemptions. The duty applies only when the advertiser has actual knowledge of the synthetic performer's presence.

So if your agency pulls a fast one on you, you’re good? Exemptions cover audio-only ads, AI used solely for language translation of a human performer, and promotional materials for expressive works (movies, TV, streaming, video games) where the synthetic performer's use is consistent with the underlying work.

(This is legalese for: if the movie was made with AI, you can use AI in the ad, because you used AI to make the movie.) 

The law also doesn't apply to media outlets or platforms that merely publish or disseminate non-compliant ads, and it doesn't affect Section 230 protections. Liability sits with whoever produces or creates the ad.

This is interesting because it’s different than how liability is usually applied. The EU loves to sue Meta. 

Meta has already confirmed they will be labeling AI ad content. Not sure what this will look like yet, but at least on Meta, you’re covered – if this works accurately. 

What’s crazy: hours after Hochul signed this bill, the White House issued an Executive Order seeking to halt state-level AI regulation in favor of a yet-to-be-determined (no surprise) federal standard. The AI firms are lobbying DC hard, also no surprise. 

Some states may face federal challenges to their AI laws, but the Executive Order itself is likely to be challenged in court. In the face of that uncertainty, the prevailing legal advice I’m reading is that advertisers should continue preparing to comply with the New York law. 

Enforcement runs through the state, not consumers. So most law firms are suggesting the Attorney General’s office will enforce this, but I have no idea how. There’s no private right to action, so you can’t sue some other advertiser you see breaking the law. 

If you ask me, this law is designed to stop these horrific, malevolent, offensive, and psychologically manipulative dropshipping ads. And good riddance. These ads create a race to the bottom among advertisers who are decent human beings. They also erode the public’s trust in ads.

Trust is critical for our success as marketers. This law is a good thing.

Our friend Andrew Foxwell published a sublime tl;dr on how you can prepare as an advertiser. I recommend you start there.

Step 1: label all your AI ad creative now, if it features fake influencers. We’ll be covering this story as it develops.

🍃 Meta conversion rates are at a 12-month low. Breaking down this troubling (but maybe not bad?) trend.

📲 The top 1.4% of ad accounts are responsible for 36% of all the ad creative on Meta. Creative volume envy?

🤝 Taylor Holiday and Austin Harrison discuss the details of the CTC x Northbeam partnership. Making Northbeam accessible for aspirational teams.

🤑 ChatGPT conversion ads go live June 5. Computer, buy me an espresso.

🧾 OpenAI just announced more details on ChatGPT ad pricing. Recommended $3-5 CPC bids and a $60 default max CPM. Sorry, I can’t hear you over the sound of my own screaming.

🧣 How NAADAM tests into new channels. A straightforward approach from some experienced operators.

Love this newsletter? Please forward it to your friends, coworkers, or Victor Wembanyama and De’Aaron Fox, who combined were only 9-for-34 from the field in game one. Knicks in five. Subscribe here to get this data in your inbox every week.