The Case That Says It All

Francheska Pujols, a 28-year-old model and actress based in Manhattan, signed a deal with clothing retailer Rainbow USA in September 2024. Simple enough. She posed against a white background in various outfits — arms at her sides, neutral expression, nothing provocative.

Then Rainbow allegedly took those clean photos and fed them into an AI system.

What came out looked nothing like what Pujols agreed to. According to a lawsuit filed May 22 in New York Supreme Court and reported by the NY Post, finalized ads showed an AI-generated replica of Pujols in sexually suggestive poses — lying on another woman's lap holding a cocktail, straddling a barstool with her legs spread open.

She never posed that way. She never agreed to it. And the ads reportedly kept running after her contract expired on March 15.

Pujols sent a cease and desist in March. Rainbow allegedly ignored it. The lawsuit accuses the company of defamation, misappropriation of likeness, and violating her right of publicity. She's demanding a jury trial.

The Same Thing Is Happening at Other Companies

The same week Pujols filed her suit, Futurism reported on a separate case involving Superhuman.

Last August, Superhuman quietly launched a feature called "Expert Review," which let users get writing feedback from what the company marketed as AI clones of professional writers — using their names, their styles, and their reputations to sell a product. Without asking them.

Investigative journalist Julia Angwin led a class-action lawsuit. Nilay Patel, editor in chief of The Verge, was one of the cloned writers. He confronted Superhuman CEO Shishir Mehrotra directly on his podcast.

Mehrotra's response was a masterclass in corporate dodge. "It deeply pained me to feel that we under-delivered for them," he told Patel — which is not the same as saying "we were wrong" or "we'll compensate anyone."

When Patel pushed on whether the writers should be paid, Mehrotra drew a line between "attribution" and "impersonation" — arguing that using someone's name to train and sell an AI product is just attribution, not impersonation. The CEO of a company that cloned real people's professional identities to sell subscriptions was suggesting the only obligation is a name credit.

Superhuman pulled the feature in early March after the blowback. Mehrotra said the lawsuit's impersonation claims are "without merit."

What the Law Actually Says — Right Now

There is no comprehensive federal AI identity law.

Legal firm ZwillGen published an analysis on May 28, 2026, making clear that plaintiffs aren't waiting for Congress to act. They're stretching existing statutes — wiretap laws, biometric privacy acts, consumer protection codes, right-of-publicity statutes — to cover AI conduct.

The results in court are mixed. Some cases survive to trial. Others get tossed at the motion-to-dismiss stage. It depends heavily on which state you're in and which judge you draw.

New York's right-of-publicity law is actually one of the stronger ones in the country. That's likely why Pujols filed there. But even that law wasn't written with AI-generated deepfakes in mind.

Legal firm Cummings & Cummings points out another layer most people miss: AI-generated content may not even be copyrightable under current U.S. law. The Copyright Office has consistently held that works created by non-human actors don't get federal protection. That means Rainbow may not even own the AI images it allegedly created — and competitors could reuse them freely.

None of that helps Pujols. But it shows how tangled this is getting.

What Mainstream Coverage Is Getting Wrong

Most reporting on AI identity theft focuses on celebrities and tech executives — the Scarlett Johansson voice cloning story, the Taylor Swift deepfakes. High-profile enough to generate clicks.

What's getting less attention: this is happening to working-class professionals. Models, journalists, writers, customer service workers whose voices get recorded and trained on without consent. People without A-list lawyers or massive social media followings to pressure companies into backing down.

Pujols isn't a household name. She's a working model trying to protect a career she built. Rainbow USA is a budget clothing chain, not a Silicon Valley unicorn. This story doesn't fit the sexy AI narrative — but it's more representative of where the real harm is happening.

Also missing from most coverage: the expiration problem. Pujols had a contract. It expired. Rainbow allegedly kept using her likeness anyway. That's not a gray area. That's not an AI ethics debate. That's a company deciding a cease and desist letter is just noise.

What This Means for Regular People

If you've ever signed a modeling contract, appeared in a company video, had your voice recorded for a customer service call, or let a platform use your writing — you may have less control over your identity than you think.

Companies are betting that most people won't sue. Filing a lawsuit is expensive, slow, and emotionally brutal. Most people fold.

The ones who don't — like Pujols, like Angwin — are building the case law that will eventually force Congress to act. Or force companies to stop. The current rule is simple: if you didn't explicitly consent to AI use of your likeness, your voice, or your work, assume someone is already testing whether they can get away with it.

Because some of them clearly are.