AI Identity Rights

State of play.

• Trademark law has emerged as the primary defensive tool for celebrities facing AI deepfake and voice-cloning risk. Taylor Swift's company filed three USPTO applications on April 24, 2026 covering spoken phrases and a stage photograph, following Matthew McConaughey's earlier registrations for his voice, image, and catchphrase — both treating identity as a commercial source identifier under the Lanham Act .
• Facial trademark registration is gaining traction in the UK and EU. Luke Littler filed to register his face with the UK IPO in March 2026, building on footballer Cole Palmer's successful photorealistic facial mark registered in November 2025; the EU Grand Board of Appeal has a pending case (Jan Smit) that will set the standard for photorealistic facial marks across the bloc .
• New York has enacted consent and disclosure mandates for synthetic performers in fashion and beauty advertising, effective June 19, 2026, with California's parallel statutes already in force and a federal NO FAKES Act pending — creating a fragmented, multi-regime compliance environment .
• Voice cloning has shifted from a product feature to an active liability vector, with class action activity and settlement exposure now documented across the entertainment and gaming sectors .
• For counsel advising entertainment, sports, media, or fashion clients, the practical baseline is that identity protection now requires a layered strategy — trademark filings for commercial source-identifier coverage, right-of-publicity and consent protocols for synthetic replicas, and immediate compliance audit against New York's June 2026 effective date.

Where things stand.

• Trademark as the anti-deepfake vehicle of choice. The Swift and McConaughey filings treat voice, likeness, and catchphrases as Lanham Act source identifiers — a theory that targets commercial misuse implying endorsement, filling a gap copyright cannot reach when no specific work is copied .
• Facial trademark doctrine is unsettled but moving. The UK IPO's acceptance of Cole Palmer's photorealistic likeness mark signals openness to generic representations of well-known figures; the EU Grand Board of Appeal's Jan Smit decision will be the authoritative statement on whether photorealistic faces meet distinctiveness requirements .
• State consent regimes are the operative law for synthetic performers. New York's Fashion Workers Act (S9832) and synthetic performer disclosure laws (S.8420-A/A.8887-B), signed December 11, 2025 and effective June 19, 2026, require explicit consent before digital replication and mandate AI-avatar disclaimers in advertising; California's AB 2602/AB 1836 are already in force .
• Federal preemption risk is live. A White House Executive Order issued December 11, 2025 seeks federal preemption of conflicting state AI laws, creating a potential collision with New York's and California's consent mandates; the NO FAKES Act remains pending .
• EU AI Act disclosure obligations layer on top of state regimes. Effective August 2026, the EU AI Act requires labeling of AI-altered content with penalties reaching €15 million — relevant for any brand with EU-facing advertising .
• Voice cloning litigation is producing settlement data. Class action filings and settlements in the entertainment and gaming sectors are establishing a damages baseline for unauthorized voice synthesis .
• The USPTO has no settled standard for identity-based sound or likeness marks. The scope of protection for voice clips and photographic likenesses as trademarks remains largely untested in litigation; the Swift applications are the leading test case .

Latest developments.

• Taylor Swift's company filed three USPTO trademark applications on April 24, 2026 — two audio clips of her spoken phrases and one stage photograph — as an anti-deepfake measure .
• Reporting confirms Swift's broader trademark strategy, with Venable LLP filing under the Lanham Act to treat vocal patterns and introductory phrases as commercial source identifiers, following McConaughey's precedent .
• Luke Littler filed to register his face as a UK trademark in March 2026 across gaming and entertainment categories; the EU Grand Board of Appeal's Jan Smit case is pending .
• New York's Fashion Workers Act and synthetic performer disclosure laws take effect June 19, 2026, requiring consent protocols and AI-avatar disclaimers .
• Voice cloning class action and settlement activity is documented across entertainment and gaming, establishing early damages benchmarks .

Active questions and open splits.

• Whether the USPTO will register voice clips and spoken phrases as source identifiers. The Swift applications are the first significant test; if approved, they establish a new defensive layer for any high-profile client — if narrowly construed or refused, the gap in anti-deepfake protection widens and right-of-publicity claims become the fallback .
• Whether photorealistic facial likenesses meet trademark distinctiveness requirements. The UK IPO has signaled openness via Cole Palmer; the EU Grand Board of Appeal's Jan Smit ruling will either harmonize or fracture the transatlantic standard — a direct advisory question for clients with cross-border identity portfolios .
• Federal preemption vs. state consent regimes. The December 2025 Executive Order's preemption posture is on a collision course with New York's and California's consent mandates; how courts resolve the conflict determines whether a single federal compliance framework emerges or brands must manage a 50-state patchwork .
• Trademark infringement doctrine when the infringing content is AI-generated. Courts have not yet addressed whether AI-generated deepfakes that imply celebrity endorsement satisfy the likelihood-of-confusion standard — the answer shapes both enforcement strategy and platform liability exposure .
• Voice cloning liability allocation between developers, platforms, and end users. As settlement data accumulates, the question of where liability lands — at the model-training layer, the API layer, or the end-user deployment layer — remains unresolved and is the central drafting question in any AI voice product agreement .
• Whether synthetic performers in advertising constitute deceptive trade practices independent of consent violations. New York's regime focuses on consent and disclosure, but FTC authority over deceptive endorsements could produce a parallel enforcement vector that survives any preemption of state consent laws .

What to watch.

• USPTO examination actions on the Swift voice and likeness applications — first office actions will signal whether the examiner treats spoken phrases as inherently distinctive or requires acquired distinctiveness evidence.
• EU Grand Board of Appeal ruling in the Jan Smit case — the authoritative statement on photorealistic facial marks that will directly affect cross-border identity portfolio strategy.
• New York Department of Labor model agency registration deadline and first enforcement actions under the Fashion Workers Act and synthetic performer disclosure laws (effective June 19, 2026).
• Whether the NO FAKES Act advances and how its federal right-of-publicity framework interacts with the preemption posture of the December 2025 Executive Order.
• EU AI Act AI-content labeling enforcement beginning August 2026 — the first penalties will calibrate how aggressively the regime applies to fashion and entertainment advertising.