Can I be sued if my AI fashion model resembles a celebrity or influencer?

Quick Answer

Yes, you can be sued, but a resemblance alone is not automatically unlawful. Risk rises when the image is recognizable as a particular real person and is used commercially without consent, or when the ad is likely to make consumers believe that person sponsored or approved the product. Right-of-publicity law varies by state, while federal trademark and advertising rules can also apply.

When does resemblance become an identifiable use of a real person?

A generic face that merely shares common features with a public figure is different from appropriating that person's identity. The exact test depends on the state, but recognizability and commercial use are central. For example, California Civil Code Section 3344 covers knowing use of another person's name, voice, signature, photograph, or likeness in products or advertising without consent, and defines a person in a photograph as readily identifiable when a viewer can reasonably determine who is depicted.

An exact facial duplicate is not required in every claim. Distinctive attributes, captions, the prompt or source images, styling, the target audience, and the product category can make a particular person's identity clear. In false-endorsement cases, courts also examine whether consumers are likely to be confused about sponsorship or approval, not similarity in isolation. A 2025 Ninth Circuit decision restates that public figures may bring Section 43(a) claims over a persona, likeness, or other uniquely distinguishing characteristic when its use is likely to cause that confusion.

Which US laws may apply to a celebrity-resembling AI fashion model?

• State right of publicity: State rules differ on protected traits, commercial-use requirements, defenses, remedies, postmortem rights, and which jurisdiction's law applies. Tennessee's ELVIS Act, effective since July 1, 2024, expressly protects an individual's likeness and simulated voice. California's AB 1836, effective January 1, 2025, is narrower than a blanket AI-likeness law: it expanded the state's postmortem statute for digital replicas of deceased personalities in expressive audiovisual works and sound recordings.
• Federal false endorsement: Section 43(a) of the Lanham Act can apply when a person's identity is used in commerce in a way likely to confuse consumers about affiliation, sponsorship, or approval. Calling an image AI-generated does not supply consent or resolve that confusion.
• Deceptive advertising: The FTC's guidance on AI avatars says there is no blanket ban on AI avatars, but an unauthorized celebrity avatar violates the Consumer Reviews and Testimonials Rule if reasonable consumers would believe the celebrity gave the testimonial. An avatar can also be deceptive under the FTC Act.
• AI disclosure: Since June 9, 2026, New York General Business Law Section 396-b requires a conspicuous disclosure when an advertiser knowingly uses a synthetic performer, defined as a generated performer who is not recognizable as an identifiable natural performer. That transparency rule applies to generic synthetic people; it does not authorize copying an identifiable celebrity or cure a misleading endorsement.

As of August 19, 2026, the federal NO FAKES Act of 2025 remains introduced rather than enacted. There is still no single federal right-of-publicity statute replacing the state-by-state analysis.

How can a brand reduce likeness and false-endorsement risk?

• Avoid celebrity names in prompts and Custom Directions (the user-written instructions layered on top of selected ingredients), including phrasings like "in the style of [Name]" or "[Name]'s face."
• Do not use a real person's photos to create a custom identity unless written consent expressly covers AI or digital-replica use, media, territory, term, and compensation.
• Run a blind recognition check before publication. Ask reviewers whether the person reminds them of anyone without suggesting a name. Repeated identification of the same person is a warning to stop and obtain legal review.
• Treat disclosure and consent as separate questions. An AI label may satisfy a transparency rule, but it does not grant likeness rights or remove a false implication of endorsement.
• Keep the source images, consent, generation record, final Asset, and publication channels together for the campaign.

Nightjar calls its reusable AI people Fashion Models. A Team can choose from 80+ pre-built Fashion Models or create a custom Fashion Model from one to five source Assets when it holds the necessary rights. Saving the selected Fashion Model in a Recipe, Nightjar's reusable Create-form setup, helps the Team reuse that identity across later Generations instead of asking the system to invent a new face each time. Every output still needs review before publication.

This article provides general US information, not legal advice. For a specific campaign, ask a qualified advertising or IP lawyer which jurisdiction's law applies and whether consent, disclosure, or both are required.