Digital models are coming for fashion’s creative industry, regardless of consumer criticisms.
In late March, Swedish fast fashion giant H&M announced it would be creating artificial intelligence-generated “digital twins” of models to be used in marketing campaigns. The models will own the rights to their own AI image, which can also be used by brands not associated with H&M. “We are curious to explore how to showcase our fashion in new creative ways — and embrace the benefits of new technology — while staying true to our commitment to personal style,” says Jörgen Andersson, chief creative officer at H&M.
The announcement was met with backlash, echoing the response to Levi’s 2023 statement that it would use AI to “supplement” human models. People are clearly uncomfortable with the idea of AI models, often citing concerns around job displacement, creative authenticity and consent. But the fashion industry is drawn by the promises of speed, cost-efficiency and limitless customisation. Mango followed Levi’s lead with a fully AI-generated campaign for its youth line last July, while Coach, Estée Lauder, Hugo Boss, Moncler and Revolve have each begun experimenting with the technology. H&M’s move signals AI’s growing role in high-profile campaigns — and its creeping normalisation.
Many saw the news, coupled with the release of OpenAI’s sophisticated new image generator that allows any photo to be redone in the style of (among others) Japanese filmmaker Hayao Miyazaki and his famed Studio Ghibli, as a watershed moment. The tool revealed just how rapidly and convincingly AI can mimic highly stylised, recognisable aesthetics, making it easy to imagine a future where entire fashion campaigns, lookbooks, or brand worlds can be conjured without a single photoshoot. It raised urgent questions about creative ownership, authenticity and what role human artistry will play in an industry increasingly shaped by synthetic visuals.
An unstoppable shift
AI can already produce hyperrealistic models, often trained on imagery of real fashion models — frequently without consent. This is one of the core ethical flashpoints: AI tools often scrape public imagery without input or consent, detaching models from their own likeness and labour.
“We can easily create digital models of any type, any ethnicity, any body type — anything you want,” says Cyril Foiret, founder of Maison Meta, a generative AI agency that has worked with the likes of H&M, Mango and Revolve.
“In many ways, the biggest problem I have with what’s happening and the kind of pathway to augmentation and automation is that fashion models’ images were already used to train these AI models without their consent,” says model and futurist Sinead Bovell, who predicted in 2020 that AI would eventually replace models’ jobs. “That has already taken away a lot of power from fashion models. This technology is built on an entire chain of exploitation of modelling and of the industry.”
Bovell says AI’s infiltration into the fashion industry is happening faster than initially predicted. “In e-commerce, we’re maybe 95 per cent there in the sense that the public likely can’t tell if it’s a human or AI model. In high fashion and editorial, there’s still a way to go — those shoots rely on emotional storytelling from real models,” she says. But in time, with a good creative director, perhaps, who can refine the prompts, they likely will eventually get an AI-generated fashion model to at least convey some emotional cues that could resonate with the average viewer, or with the average person that’s flipping through a magazine or scrolling online.”
If that happens, the implications are profound: modelling could become less about physical presence and performance, and more about licensing one’s image. It opens up new strategies for brands — faster content creation, global scalability and reduced production costs. But for models and creatives, it raises existential concerns about job security and what it means to be ‘cast’ in a campaign when your likeness can be generated with a prompt.
“Generative AI is being used in an industry uniquely prone to exploitative practices with few legal protections in place,” says Model Alliance founder Sara Ziff. A 2023 poll by the organisation found that 87 per cent of surveyed models and influencers believed AI-generated models had a mostly negative impact on their industry. “We want to be clear that we are not anti-tech — indeed, we see promise in technology’s ability to improve the fashion industry — but we have serious concerns about the use of digital replicas without meaningful protections in place.”
Models depend on their appearances to make money, but as AI tools enable easier photo manipulation, the need for laws and guidelines is coming into sharp relief.
H&M’s digital twin initiative attempts to address concerns around consent by giving models rights over their AI likeness — an approach positioned as ethical, but still controversial. “H&M’s initiative could have significant and long-standing repercussions for creative workers throughout the fashion industry,” says Ziff. “It raises critical questions about consent and compensation: who will legally own the digital twin, the model or the agency? And which one will clients negotiate contracts with? How will models be fairly compensated for their digital twin? Will rates increase because the company is saving on travel costs, or be reduced since models won’t have a physical presence? What are the structures for payment, including day rates?”
The shift could also mean significant job losses, impacting the livelihoods of a host of fashion workers, from photographers and stylists, to makeup artists, hair stylists and caterers.
Despite internet backlash, public opinion is split. In Sprout Social’s recent survey of 2,000 UK consumers, half said they would be comfortable with a brand using AI influencers in addition to human influencers. “Where brands will run into the biggest problem is if a model is supposed to represent the experience they are modelling or tell an emotional story about how an item made them feel, or how effective skincare was,” says Bovell. “I don’t think AI should speak on behalf of experiences it hasn’t gone through. That is strange and in some ways quite deceptive.”
The case for digital twins
Foiret of Maison Meta argues that digital twins could actually help models retain agency and income. “We encouraged agencies like Next to digitise their model catalogues to stay relevant,” he says. “Otherwise, they’ll be replaced.” Foiret adds that “every” modelling agency will eventually adapt.
“While we understand that this will spark opinions, discussions and uncertainty, AI will be part of many processes going forward, and we are humble in acknowledging that we do not hold all the answers at this point,” says H&M’s Andersson. “We will continue to explore with other creatives within the industry to see what generative AI can bring to any creative process, while ensuring we maintain our human-centric approach.”
Luxury houses will likely preserve human models for campaigns, but for e-commerce, social media and marketing visuals, a hybrid approach is likely. “L’Oréal told us they’ll always use humans in campaigns across their brands — but AI could support background elements or on-screen graphics,” says Foiret.
Bovell believes digital twins could be a fair compromise if the correct frameworks are implemented. “This way, models are still being compensated for their time and their talent,” she says. It also can be a way of preventing brands from using AI to generate ‘diverse’ models without involving real people.
Brands experimenting with AI are finding ways to represent people from marginalised groups in digital form — with mixed responses from consumers and industry insiders. Here’s how to get it right.
“There are all sorts of nuanced ways for exploitation and misrepresentation when that happens,” she says. “As a society, we have to decide where the new lines are in terms of cultural appropriation that we want to draw in the digital world.”
Foiret predicts consumers will soon see AI-generated versions of themselves modelling clothes, creating more personalised, interactive experiences. The shift could also help brands cut costs, save time and reduce e-commerce returns, which contribute millions of tonnes of carbon dioxide emissions annually.
A spate of recent partnerships are illustrating how NFTs can more closely integrate with physical goods.
Still, as AI tools become further integrated into the creative workflow, questions of ethics and regulation remain critical. Brands using AI-generated imagery should be “clear and transparent”, says Foiret. The Fashion Workers Act, a New York State bill that goes into effect in June, will require modelling agencies and brands that wish to create or use a model’s digital replica to obtain the individual’s clear, written consent while detailing the scope, purpose, rate of pay and duration of the use in order to create or use their digital replica. “There’s no need for a brand to wait to begin implementing these practices,” says Ziff.
Companies should consider how they can harness AI technology to supplement the work of human models and other fashion creatives, but this requires having meaningful protections in place. “Models’ abilities to use their digital replicas in a way that serves them hinges on consent, transparency and fair compensation,” Ziff says. “Otherwise, this technology is poised to be used in a way that is highly extractive and implemented at the expense of a wide swathe of working people.”
Comments, questions or feedback? Email us at feedback@voguebusiness.com.
