The Controversy Surrounding AI Models in Fashion

Change is rippling through the fashion industry as artificial intelligence models continue to grow in prominence. In 2023, Sarah Murray, a commercial model, recalled her initial encounter with an AI-generated model—a stunning young woman of color wearing a Levi’s denim overall dress. This was another major step forward in the growing convergence between technology and…

Lisa Wong Avatar

By

The Controversy Surrounding AI Models in Fashion

Change is rippling through the fashion industry as artificial intelligence models continue to grow in prominence. In 2023, Sarah Murray, a commercial model, recalled her initial encounter with an AI-generated model—a stunning young woman of color wearing a Levi’s denim overall dress. This was another major step forward in the growing convergence between technology and the fashion industry. Brands such as H&M, Mango and Calvin Klein are already adopting AI models to fuel their entire programmatic advertising campaigns. This change further raises the important discussion around ethics and authenticity.

Paul Mouginot, a fashion analyst, highlighted that French retailer Veepee has utilized virtual mannequins for clothing sales since at least 2013. This trend has deceptively paved the way for massive, unregulated acceptance of AI within fashion advertising. It has sparked a larger conversation around what this means for the future of human models and the authenticity of representation. What better way for brands to save money and increase operational effectiveness? In consequence, the debate on the impact of human models on the industry is becoming louder.

Amidst this backdrop, Sara Ziff, a former model and founder of the Model Alliance, is advocating for the Fashion Workers Act. This long-awaited, proposed legislation would require express consent from all models and talent. Notably, it requires that brands be compensated for the use of their digital likenesses. As we increasingly turn to AI models to answer our questions, we need to be aware of representation issues. We must equally defend the rights of human artists, too.

The Rise of AI Models

While brands are still figuring out how best to utilize AI technology, the fashion world is evolving at lightning speed. To save money, companies have started to use AI models as a cheaper alternative to create content at high volume. As fashion journalist Amy Odell explains, “It’s more cost-effective for brands to work with AI models today. Brands want more content than ever, and it simply compounds. This ongoing trend to automate has raised a lot of eyebrows on both sides among industry insiders.

Paul Mouginot explained the financial rationale behind this trend: “It’s simply expensive to work with live models, especially when it comes to photographing them in countless garments, shoes, and accessories.” The recent increased drive for supply chain and logistics efficiency and cost savings has many brands going all in on AI without any thought for traditional modeling practices. Too frequently, this decision is made at the detriment of human touch and authenticity.

Then there’s Sinead Bovell, who speaks about ethical representation in fashion and criticized the trend as “robot cultural appropriation.” She emphasized that e-commerce models are particularly vulnerable to automation and that the industry must prioritize telling unique human stories rather than relying on AI-generated representations. “AI will never have a unique human story,” she asserted.

Ethical Considerations in Fashion

As the popularity of AI-generated models has surged, so too have conversations about ethics in fashion advertising. Murray particularly lamented the ramifications of such a shift. “If those brands ever had the opportunity to stand in line at an open casting call, they would know about the endless amounts of models, including myself, that would dream of opportunities to work with their brands,” she said. This sentiment resonates with many who believe that relying on AI undermines the potential for real human connection in fashion.

Ziff’s efforts to pass the Fashion Workers Act are indicative of a larger movement in the industry that recognizes these concerns and more. The Act’s purpose is to address issues of consent and compensation for use of digital replicas. It signals a commitment to ethical practices as the technology evolves. As Ziff’s advocacy makes clear, we must ensure there’s accountability in this industry that is becoming ever more dependent on automation.

PJ Pereira, co-founder of AI ad agency Silverside AI, sent out that warning. This will likely create bias against traditional, non-AI generated fashion advertising, as added by Senator Hawley. He stated, “As more content gets made by more people who aren’t intentional, all of the output feeds back into computer models, amplifying bias.” This underscores the need for deeper and more intentional engagement with developing technology as brands continue to balance the needs of diverse representation against the perils of inauthenticity.

The Future of Fashion and Technology

The collision of fashion and technology is one that will intensify and expand as AI becomes further embedded into advertising strategies. Sandrine Decorde, Chief Executive Officer and co-founder of Artcare, added that her company is committed to designing human-centered AI models. “The more time we spend on our datasets and image refinements, the better and more consistent our models are,” she explained. This dedication to improving AI design raises questions about the future role of human models in an increasingly automated industry.

At a press conference for the show, Murray shared her discontent with the direction fashion advertising seems to be headed. “They would never need to supplement with anything fake,” she argued. Her viewpoint highlights the artistic nature that human models add to fashion storytelling—an element that AI will never be able to achieve.

Pereira admitted that scaling from today’s modeling practices to an in­-house, fully AI-driven approach would need more than changes to processes. “There’s no way to scale from four to 400 or 400,000 with just process tweaks,” he stated. This recognition marks a maturation of thinking about the realities and challenges of moving toward an AI-driven paradigm.