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

Quick Answer

Yes. Generating an AI image that resembles an identifiable real person without permission can violate their right of publicity, and using it to sell a product can also trigger false-endorsement claims under Section 43(a) of the Lanham Act. The image being AI-generated is not a defense, and recent state laws (the Tennessee ELVIS Act, California AB 1836, and the New York synthetic-performer disclosure law taking effect June 9, 2026) make the surface broader, not narrower. A federal NO FAKES Act has been reintroduced in the 119th Congress but has not passed either chamber as of May 2026.

The legal risks

• Right of publicity: Using someone's identifiable likeness for commercial gain without consent can expose you to damages. The right is governed state by state, can cover face, distinctive look, voice, and signature mannerisms, and several states recognise post-mortem rights. California's AB 1836, in effect since January 1, 2026, extends California's post-mortem statute to AI digital replicas of deceased performers. Tennessee's ELVIS Act, in force since July 1, 2024, expands the Tennessee statute to cover voice and AI-generated likeness whether or not the subject lives in the state.
• False endorsement: If the image suggests a celebrity endorses your product when they do not, you can face Section 43(a) Lanham Act claims for misleading advertising, in addition to any state right-of-publicity claim.
• New York synthetic-performer disclosure: From June 9, 2026, advertisements distributed in New York that feature an AI-generated person require a conspicuous AI disclosure under amended General Business Law Section 396-b. Penalties are USD 1,000 for a first violation and USD 5,000 per subsequent violation. This is in addition to the right-of-publicity rules above.
• Federal: A federal NO FAKES Act has been reintroduced in the 119th Congress with bipartisan sponsorship but has not been enacted as of May 2026. Until it is, the surface is state law plus the Lanham Act, not a single federal right of publicity.
• Indemnification: Most generic AI tools state in their terms that the user is responsible for inputs and outputs. If you prompt for a celebrity, the liability lands on you, not the tool.

How Nightjar reduces this risk

Nightjar separates the person in the shot from the prompt by giving every Team a roster of reusable AI people called Fashion Models: a Team can pick from 80+ pre-built Fashion Models or create a custom Fashion Model from their own source Assets, controlling age range and gender presentation without prompting for a famous face. A custom Fashion Model based on a real person should only be created when you have the right to use that person's likeness. Saving an approved Fashion Model into a Recipe (Nightjar's saved Create-form setup, which captures the photography style, composition, model choice, background, and output settings) keeps the same vetted person across future Generations, instead of regenerating from scratch and risking a celebrity-resembling output each time.

Practical guidance

• 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."
• Use synthetic Fashion Models that are not designed to imitate a specific real person.
• Save the approved Fashion Model into a Recipe and reuse it across Generations rather than regenerating the person each time.
• Review outputs before publishing. If a Generation reads as a recognisable real person, discard it and regenerate.
• Keep an audit trail in your Library: which Fashion Model, which Recipe, and which Generation produced the Asset that ran in a given campaign. The Asset detail view records the inputs and ingredients used, which helps if you need to defend a likeness challenge.

This article is general information, not legal advice. For a specific campaign, consult a qualified IP attorney in your jurisdiction.