Authors
Introduction
AI adoption in fashion and beauty has moved from experimentation to widespread commercial use, with applications ranging from creative production and virtual try-ons to synthetic models and personalized experiences. The key legal question is no longer whether AI can be used, but whether its use is transparent, authorized, and compliant with existing rules on personality rights, data protection, advertising, and consumer protection.
Industry practice: The rise of AI-generated models and digital replicas
AI-generated models and digital replicas are increasingly common. Mango launched an AI-generated campaign across 95 markets in 2024, while H&M announced in 2025 the creation of AI “digital twins” of 30 models, raising key issues around consent, personality rights (right of publicity in the United States), and disclosure. H&M indicated that models retain rights in their digital twins, are paid similarly to current arrangements, and that AI-generated content is disclosed through visible watermarks.
In response to these developments, specific laws addressing AI-related uses of models’ identities have emerged. New York’s Fashion Workers Act requires clear, prior written consent from a model before creating or using a digital replica of their face, body, or voice that specifies the scope, purpose, duration, and compensation. Such consent must be clear and conspicuous and must be separate from standard representation agreements. Notably, the Fashion Workers Act excludes standard or traditional post-production editing (e.g., coloring, retouching). Separately, New York’s Synthetic Performer Disclosure Law requires a clear disclosure when advertisements include a “synthetic performer” – a digitally created asset intended to create the impression of a human performer who is not identifiable as any real person. This covers TV, digital, and social media advertising but excludes audio-only ads and expressive works.
At the federal level, at the time of this writing, the proposed bipartisan NO FAKES Act, if passed, would create a federal IP right to authorize use of an individual’s voice or visual likeness in a digital replica or via a service or product designed to produce unauthorized digital replicas of specific individuals, with limited exceptions.
SAG-AFTRA 2025 Commercials Contract: Digital replica and synthetic performer provisions
In April 2025, SAG-AFTRA ratified a new three-year Commercials Contract introducing consent and compensation guardrails for AI. The contract requires 48 hours’ notice and informed consent before creating or using a “digital replica” of a union member – a computer program made using a performer’s voice, image, or performance that can independently generate new performances. Compensation for digital replica use triggers 1.5x scale session fees plus holding and use fees at scale to the performer. Similar to the Fashion Workers Act, traditional post-production and visual effects (VFX) methodologies are not included.
The contract also addresses “synthetic performers” – digitally created assets not identifiable as real people. Synthetic performers may only be used in commercials that also include a human principal performer. Provided there is one human principal performer to cover the synthetic performer, the producer must pay the equivalent of what a human performer would have received into the SAG-AFTRA Pension and Health Plans. Additionally, performer services under the contract cannot be used to train generative AI without union consent. For fashion and beauty brands engaging union talent, these provisions layer consent, compensation, and data-use obligations on top of applicable state law when using AI-generated or AI-enhanced performances.
AI regulatory framework
The United States still does not have a comprehensive federal AI framework. Fashion and beauty companies must therefore rely on existing legal frameworks, including consumer protection rules, privacy laws, publicity rights, and contractual protections.
The EU AI Act (Regulation (EU) 2024/1689) establishes the EU’s main AI regulatory framework. Article 50 of the EU AI Act introduces transparency obligations that apply to certain AI systems (e.g., AI systems that interact with humans such as chatbots, generative AI, and emotion recognition AI). In fashion and beauty, most applications, including AI-generated campaigns, chatbots, and synthetic imagery, will fall under those obligations.
Article 50 of the EU AI Act applies as of August 2, 2026, and requires informing users when they interact with AI systems and ensuring that certain AI-generated content is appropriately labeled. These obligations also cover deepfakes and synthetic media, subject to limited exceptions. Non-compliance may result in fines of up to €15 million or 3% of worldwide annual turnover.
The Digital Omnibus provided a limited transitional period for providers of AI systems, including general-purpose AI systems, that generate audio, image, video, and text content and were placed on the market before August 2, 2026, requiring them to comply with the AI Act’s Article 50(2) labeling obligations by December 2, 2026.
Article 5 of the AI Act, as amended by the Digital Omnibus, introduces additional prohibitions on the placing on the market, putting into service, or use of AI systems that generate or manipulate non-consensual intimate material or child sexual abuse material, effective December 2, 2026.
In June 2026, the European Commission published a voluntary Code of Practice on marking and labeling AI-generated content, providing practical guidance for providers and users of generative AI systems. Moreover, optional EU icons may be used to identify certain AI-generated content but would not, on their own, establish compliance.
By way of example, certain EU Member States have already adopted transparency requirements for AI-generated content. For instance, France introduced similar requirements through the 2023 Influencer Act (Law n° 2023-451), which requires influencers to label digitally altered images and AI-generated images used in commercial content.