Does Your Image License Cover AI? Pujols v. Rainbow USA and the Risks of AI-Generated Model Replicas

A recently filed lawsuit and New York’s Fashion Workers Act suggest the answer may be no. The legal takeaway is straightforward: standard photo-use rights should not be treated as consent to automatically authorize AI-generated replicas unless the agreement expressly authorizes that use. For brands, this issue is becoming increasingly urgent as AI tools are integrated into workflows, creating consent risk when a real model’s likeness is involved. For brands that get ahead of this issue, proactive compliance is an opportunity to build stronger talent relationships and more defensible content practices.

The Dispute

Model Francheska Pujols filed suit against apparel retailer Rainbow Shops in New York Supreme Court (Pujols v. Rainbow USA, Inc., Index No. 157543/2026). The lawsuit alleges that Rainbow took photographs from standard commercial photoshoots and used AI to create entirely new images showing the model in different poses and settings, including depictions she claims she did not authorize, without obtaining her consent. Rainbow has denied wrongdoing, and the case remains in its early stages.

Regardless of how this case is ultimately resolved, it highlights the type of conduct the Fashion Workers Act is designed to address: using AI to create and commercially exploit digital replicas of a model’s likeness without specific authorization. The compliance path forward is clear. Most existing talent agreements were written before AI image generation became commonplace and updating them now can help brands close the gap between what those contracts authorize and what AI makes possible.

The Central Legal Question

New York’s Fashion Workers Act (NYS Labor Law, Article 36), which took effect on June 19, 2025, draws a clear line: there is a meaningful difference between editing a photograph and using AI to generate an entirely new one.

What Is a ‘Digital Replica’?

The Act defines a “digital replica” as a computer-generated or AI-enhanced representation of a model’s likeness, including their face, body, or voice. Importantly, standard post-production work remains unaffected: color correction, blemish retouching, and background cleanup are still permissible under existing agreements. The line is crossed only when AI is used to create something materially new. A useful rule of thumb is if the brand would traditionally have needed the model to return to the studio to create the image (a new pose, a different outfit, an alternate setting, or placement alongside a new product), then express consent is required for AI generation.

Dual Consent Obligations

The Act places separate obligations on both brands and modeling agencies. Brands must obtain “clear and conspicuous prior written consent” that specifies the scope, purpose, rate of pay, and duration of any AI use. Agencies must obtain their own separate written consent apart from the standard representation agreement. While this dual requirement adds a step, it also protects brands: proper documentation creates a clear record that consent was obtained, which can be valuable if questions arise later.

Why Traditional Licenses Fall Short

Traditional image agreements typically covered media channels, geographic territory, campaign length, and reuse rights. They were not written with AI generation in mind, and there was no reason they would have been. The Act now requires that AI consent be specific: it must detail scope, purpose, rate of pay, and duration. Broad, catch-all language in older contracts (“all media now known or hereafter developed”) may not be enough, but this is a straightforward gap to close with updated consent forms.

Liability Beyond the Act

The Fashion Workers Act is not the only consideration. Brands should think about AI-generated likeness risk in several distinct categories. 

1. There is statutory risk under the Act if a digital replica is created or used without clear, conspicuous, and specific prior written consent. 

2. There is contract risk if the use exceeds the scope of the model release, agency agreement, or campaign authorization. 

3. There may be right-of-publicity or false endorsement exposure if the generated image makes the model appear to endorse a product, campaign, or context they never approved. 

4. Consumer protection and reputational risks may arise if AI use is not transparent or appears deceptive. At the same time, brands with documented AI consent protocols are better positioned to show that they exercised reasonable care if questions arise later.

Key Takeaways

• Audit and update talent agreements. Review whether existing provisions clearly and conspicuously cover AI-generated digital replicas and specify scope, purpose, rate of pay, and duration. Agreements that do not may warrant review before any digital replica is created.

• Use standalone AI consent forms. AI consent should not be buried in general image-use provisions or standard boilerplate. Separate consent forms can help address the Act’s requirements and clearly delineate what is permitted: routine retouching, AI-assisted editing, material alterations, and fully AI-generated replicas. Where there is any ambiguity,  greater specificity may help reduce risk.

• Specify asset duration. Address whether AI-generated assets survive the underlying campaign and stipulate a term for rights. Open-ended authorizations invite disputes.

• Implement clearance processes. Before any image is published, brands may want to consider whether their workflow includes a step to confirm: was this asset edited, or was it generated? If generated, is specific consent in place? This consideration may be particularly important when multiple teams or outside vendors are involved in production.

• Manage vendor risk. An internal AI policy only protects internal workflows. Brands that acquire images through agencies, production companies, or other vendors may inherit risk if those vendors never obtained proper permissions from the model. Vendor contracts may be evaluated to determine whether they adequately address AI-related risks, including consent verification and allocation of responsibility.

• Document compliance efforts. Beyond the Fashion Workers Act, brands face potential exposure under false endorsement, publicity rights, consumer protection, and intellectual property theories. Documented good-faith compliance efforts can demonstrate reasonable care and may mitigate exposure across all fronts.

Brands that move early on AI consent protocols may be better positioned not only to manage legal risk, but also to attract top talent who increasingly value transparency around how their likenesses will be used. In a competitive market, a reputation for responsible AI practices can be a meaningful differentiator.