If you model, design, or produce content for a brand, read your contract's edit clause again this week. Two lawsuits now working through courts in New York and California, and a Bloomberg Law report published two days ago, all turn on the same small piece of language: the word that lets a company "edit," "use," or "style" your image. Brands are reading that word as permission to hand your likeness to a generative model and manufacture scenes you never shot. Creators are learning, after the fact, that the contract they signed for a product photo may have also signed away control of their own face.

The claim

Bloomberg Law reported this week that brand contracts have long given companies latitude to edit and reuse content, but that creators may now unwittingly sign away the rights to their likenesses as brands bury expansive clauses allowing them to manipulate creators' content using AI. The reporting centers on Francheska Pujols, a New York model who accused clothing brand Rainbow Shops of breaching her modeling contract by creating AI images of poses she never struck, including one showing her straddling a barstool. Pujols first sued in May, withdrew the case to pursue a private settlement, and then refiled at the New York State Supreme Court on June 15 after talks failed, this time with new evidence, an Instagram post from an affiliated storefront showing her in a pose exposing her undergarments that she says she never approved. Her original contract, according to reporting, allowed Rainbow to make minor edits like cropping and stylistic tweaks, not to generate wholly new images. A parallel case out of California tells a similar story with higher stakes. Molly Tranchin, an influencer known as FashionVeggie with more than 400,000 Instagram followers, sued underwear brand EBY Inc. in the Northern District of California on June 9, alleging the company altered a modest video she provided under contract into one exposing her breasts through a sheer bra and posted it to its own account. Her complaint alleges the imagery constituted non-consensual pornography, and advances claims under California's non-consensual deepfake statute, as well as causes of action for defamation, copyright infringement, breach of contract, tortious interference, violations of her right of publicity, intentional infliction of emotional distress, non-consensual pornography, and negligence, among others. Her contract, notably, gave her the right to review and approve any edits and stated she retained full ownership of the content, protections that did not stop the alleged alteration from happening anyway.

Why it matters

For independent models and creators, these cases confirm something that has been circulating as anecdote for a year: the gap between what a contract says and what a brand does with AI is where the real risk lives. Attorneys quoted by Bloomberg Law describe rewriting influencer agreements clause by clause because vague verbs like "use" no longer carry a stable meaning once generative tools are in the workflow. One attorney said he now deletes the word "use" from contracts entirely because "there are no magic words with AI," and "use" could mean anything. That is not paranoia, it is a direct response to what happened to Pujols and Tranchin, where existing language, written before anyone anticipated AI-generated doppelgangers, was stretched to cover conduct the model or influencer never contemplated. The regulatory backdrop is shifting too, though unevenly. New York's Synthetic Performer Disclosure Law, effective June 9, 2026, now requires advertisers to conspicuously disclose when an ad features a fabricated human likeness that does not depict a real person, with civil penalties of $1,000 for a first violation and $5,000 for each one after. Separately, New York's Fashion Workers Act, which prohibits model management companies from creating or altering a model's digital likeness through AI without clear, conspicuous and separate written consent, has been in force for a year and is central to the Pujols dispute. Neither law was built for the exact fact pattern in these lawsuits, a real person's photographed likeness fed back into an AI tool to produce new, unauthorized images, which is why plaintiffs are stacking older doctrines, right of publicity, breach of contract, defamation, alongside newer deepfake statutes. That patchwork approach works, but it is expensive and uncertain, and it depends on the model finding the images in the first place.

What to watch

Watch what happens next in both cases. Pujols v. Rainbow is now on its second filing in New York State Supreme Court, and discovery, if it proceeds, could force Rainbow to disclose what tools it used and how widely the images circulated. Tranchin v. EBY is further along procedurally in the Northern District of California and tests whether a contract's approval-rights clause can be enforced against an AI alteration the brand never disclosed making. For your own contracts, the practical lesson from both cases and from the attorneys Bloomberg Law spoke with is the same: strike vague editing language, require written, use-specific consent before any AI alteration, and state plainly that generative modification of your likeness is outside the scope of any grant of rights unless you say otherwise. If you work with New York clients, confirm any ad using a synthetic performer carries the required disclosure, and if you are a model under the Fashion Workers Act, know that separate written consent for digital replicas is not optional boilerplate, it is your leverage.

Sources