Models Feel Hemmed In By AI

Using generative AI, designers can use digital photos to adjust models' features, change what they are wearing and even deploy fully digital avatars in place of humans.

AI's economic and emotional impact on fashion models was the focus of "Responsible by Design: AI, Fashion, and Creative Labor," a New York Fashion Week 2025 event hosted by the Model Alliance, a fashion industry labor rights group. The Sept. 9 event included a presentation about AI and fashion models by Zoë West, senior researcher at the ILR School's Worker Institute; Sanjay Pinto, a fellow at the Worker Institute; and Alexandra Mateescu, a researcher on the Labor Futures initiative at Data & Society Research Institute.

The three have co-authored a forthcoming paper, based on research conducted in partnership with the Model Alliance, with funding from Omidyar Network and the Ford Foundation.

"We wanted a better understanding of what is happening on the ground with the use of AI in fashion modeling and how models are experiencing it," West said.

The team has created the research snapshot "Fashion's Data Doubles: How AI is Reshaping Modeling Work" and shared key findings in a Teen Vogue article.

The researchers said real-life models and photo shoots are important to the fashion industry overall, so replacing regular models with synthetic models or digital avatars isn't a viable future, particularly in many creative or editorial contexts where design and style matter: A model can bring a unique look that enhances a brand's appeal.

The researchers combined the Worker Institute's focus on precarious work with Data & Society's recent focus on the impact of generative AI in the workplace, particularly in creative industries such as Hollywood. They said it was easier for standards around AI to be introduced into movies and television because those workers belong to unions, whereas models are classified as independent contractors and cannot unionize. "Many would say they are misclassified," West said.

Additionally, the researchers found the models' agencies largely control models' terms of work and sign contracts on their behalf, and when brands violate contract terms, a model often has little recourse.

"There are such power inequities, and such rife violation of labor protections," West said. "Models experience a lack of ownership or control over their images, with very little transparency in contracts. Models need to be empowered to enforce any standards that are put into place. Despite the glamorous image of modeling, models are largely very precarious workers."

Through a series of interviews with models, the research team discovered that AI has worsened the common problem of brands reusing photos without permission or paying the model. Now, a brand might hire a contractor to use AI to turn a single photo into a series of images featuring multiple outfits, rather than paying for a long photo shoot.

The models reported that although they have long faced the problem of their images being "Frankensteined" without permission to mix and match body parts from multiple models in one image, AI tools have made it easier for brands to dramatically edit their images for different purposes.

The interviewees said AI was being used in ways that made them feel disposable and not in control of their public image. For example, an Asian American model said that seeing a white model's face digitally superimposed onto her body felt like a "slap in the face." Another model said that turning a clothed model's image into a nude was "image violence."

Models feared that racial bias embedded in AI tools leads to images with more narrow racialized beauty norms instead of encompassing a wide range of appearances. They also believe that AI is reinforcing unhealthy beauty norms by integrating an AI-generated ideal that does not represent real bodies.

The researchers point to the New York State Fashion Workers Act, which went into effect in June 2025, as a step in the right direction, but caution that more needs to be done to establish standards to protect models' economic and emotional interests.

Tonya Engst is a writer for the ILR School.

/Public Release. This material from the originating organization/author(s) might be of the point-in-time nature, and edited for clarity, style and length. Mirage.News does not take institutional positions or sides, and all views, positions, and conclusions expressed herein are solely those of the author(s).View in full here.