
In this article (4)
New York AI Ad Law Analysis: Persona Trademarks
Key Takeaways
- Treat synthetic media consent as a paid usage right, not a routine creative approval.
- Registerable persona assets can strengthen leverage when AI imitates voices, gestures, or catchphrases.
- Sponsors should localize disclosures because New York and California do not match.
Why it matters
- ProductAd tools need disclosure workflows that capture consent, wording, and territory before synthetic creative ships.
- InvestorsPersona rights may become durable brand assets, which changes how athlete led ventures defend endorsement value.
Synthetic ads turn faces, voices, gestures, and catchphrases into contract assets, not just brand flair.
The next endorsement problem may not be a sneaker deal gone stale. It may be an ad that sounds like an athlete, moves like an athlete, and never paid the athlete. That is where the polite language of synthetic performance starts looking like a revenue leak.
Ropes & Gray’s new analysis of New York’s AI ad law is useful because it does not treat deepfake advertising as a vibes problem. It frames persona as infrastructure: voice, face, gesture, catchphrase, and the paperwork that decides who can monetize them.
The deal: disclosure is not consent
According to Ropes & Gray, New York’s Synthetic Performer Disclosure Law took effect June 9, 2026, and applies when advertisers know a synthetic performer is being used. The firm says the law defines a synthetic performer as a computer created or modified asset meant to look like a human performer in a visual or audiovisual ad, while not being recognizable as a real person. McDermott adds the comparison that California’s SB 1050 was signed by Governor Gavin Newsom on September 16, 2026, and takes effect January 1, 2027, with disclosure duties that include audio ads and specific wording. In plain contract English, the disclosure tells consumers what they are seeing, but it does not automatically tell an athlete what they are owed.
@title New York and California ad rules
@source Ropes and Gray
@source McDermott
Rule part │ New York │ California
Start │ June 9 2026 │ January 1 2027
Audio ads │ Excluded │ Included
Wording │ Not set │ Set wording
Real person │ Not covered │ Excluded
@caption Both laws make synthetic ad disclosure a campaign operations issue.
That distinction matters for athlete marketing teams. A brand can comply with a disclosure rule and still create a separate fight over likeness, endorsement implication, or trademark use. Compliance is the floor. The money sits in consent, scope, approvals, and the ability to say no before the campaign ships.
The money: persona is inventory
Ropes & Gray says generative AI can “convincingly replicate voices, faces, mannerisms,” which is lawyer language for a familiar sports business issue: somebody found a cheaper way to simulate expensive talent. The old endorsement bundle was appearance, usage rights, category exclusivity, and media buy. Synthetic media adds a new line item: whether the sponsor can generate new performances from the athlete’s identity, and who owns the resulting creative.
This is why the rise of persona trademarks is more than celebrity vanity. Ropes & Gray says celebrities and increasingly athletes are using federal trademark law, including “sound and motion mark filings,” to protect signature voices, gestures, and catchphrases from AI replication. If a catchphrase can sell a drink, introduce a shoe drop, or move NIL merch, it is not garnish. It is brand inventory, and inventory without controls tends to leave the warehouse.
The contract: sponsors need cleaner rails
McDermott says California requires disclosures that are difficult to miss, easy to understand, and presented so a reasonable consumer would notice, read, and comprehend them in the context of the ad. It also says California requires language substantially similar to “this performance features a synthetic performer” or “no human performer is depicted.” For sponsors, that means AI creative cannot be treated as a late stage production shortcut. It needs a review path before media spend starts burning.
Legal 500, summarizing the California framework, notes that SB 1050 does not apply to ads for expressive works when the synthetic performer use is consistent with the underlying work, and also excludes uses where generative AI is limited to language translation or accessibility features. Those carveouts are not a blank check for athlete campaigns. A commercial endorsement is usually selling the product first, not merely promoting an expressive work, which is why agents should separate creative convenience from commercial exploitation.
What to watch next
Ropes & Gray says pending New York legislation could extend disclosure rules to digital replicas of real individuals, which is the part athletes and creators should track closely. The current New York definition centers on a performer who is not recognizable as a real person, while McDermott says California excludes digital replicas of recognizable individuals from its synthetic performer law. That leaves a messy middle for athlete business: the more valuable the persona, the more likely the imitation is to be recognizable, and therefore handled through other contract, trademark, and likeness tools.
The practical move is boring, which is usually where the leverage lives. Athlete teams should audit what they actually own, register what can be registered, and make AI usage a separate negotiated right rather than a sentence buried under general publicity language. Sponsors should want the same clarity, because a campaign that saves on production but clouds consent is not efficient. Watch the next round of endorsement templates, NIL collective deals, and agency rider language. That is where the deepfake ad era will either get priced or get litigated.
Sources3 sources
The reporting, announcements and research the AI editor worked from. Links open the original publisher.