Inside a cavernous photography studio in north London, large-scale prints from handbag brand Misela’s latest campaign hang from a pipe grid overhead. In one, a blonde model sits on a cognac-leather sofa next to a black-and-grey tote, a classic Manhattan skyscraper scene visible through the windows behind her. Another features a black-gowned woman perched in a Venetian gondola on a misty Grand Canal. Guests pause in front of the pairings and murmur about the lighting, the settings, the bags, until someone smiles and leans in to make the big reveal: “Amazing, isn’t it? And you know, it’s all AI.”
Artificial intelligence, increasingly used in fashion when it comes to search, ecommerce and customer service, is making inroads into marketing. Models walking Coperni’s show at Paris Fashion Week last September wore Humane’s AI Pin, a wearable multi-modal device with communications, image-capturing and digital assistant applications. Brands including Etro, Moncler, Mango, Valentino, Pandora, Revolve and Misela have used generative AI to promote new collections and create conversation-starting images. More recently, AI influencers have become a flashpoint for conversations about inclusion and representation in fashion media.
The use of AI raises questions around who is profiting and what real representation looks like. “Everybody knows that AI means something important for the future,” says Etro’s creative director Marco De Vincenzo.
For the brand’s spring 2024 campaign, he worked with digital artist and prompt designer Silvia Badalotti to write text prompts that would generate otherworldly “rooms” as a backdrop for his designs. The result: “models” in greenhouses, astrological charts and pastel-terraced structures, not unlike settings from a Star Wars palace. None of them were real; the models and the settings were the product of detailed written prompts fed into an image AI platform, while the clothes were photographed and edited into the images later.
De Vincenzo finds the “infinite possibilities” in AI inspiring. “From a creative point of view, it was amazing,” he says. “It’s modifying reality, but that’s what every creative [person] tries to do. It’s our job to try to escape and experiment, every day. Sometimes reality is not enough.”
Being at the forefront of emerging tech isn’t without its challenges. When Mango used generative AI to produce a campaign to promote a collection for its Teen line, one of the trickiest aspects was “leveraging the technology in a way that stayed true to our core aesthetic”, says Jordi Álex Moreno, the brand’s chief information technology officer. “We had to ensure that the editorial quality of the final images matched our standard fashion campaigns.”
They did so by including representatives from the design, art and styling, data and photography teams on the project, rather than making it a purely digital project. “It is an excellent example of teamwork between human handcrafted intelligence and digital intelligence,” Álex notes, adding: “AI should serve as a co-pilot to amplify our employees’ capabilities and creativity, speeding up repetitive tasks so teams can spend more time on value-added work.”
Indeed, AI can create realistic imagery in minutes — seconds, even — compared to the months of shoots, reshoots and post-production typically required for a campaign. Misela’s handbag campaign, marking its 15 years in business, featured images of 15 women wearing its handbags in 15 different cities. “We wanted to take the bags and the women on a global tour,” Misela founder Serra Türker says.
End to end, the campaign took two and a half months to produce. Türker estimates it would have cost at least 15 to 20 times more to execute the campaign in a more traditional way.
Türker made mood boards for the campaign, just as she would have for a regular photo shoot. Then Alphan Eşeli, a filmmaker, screenwriter, photographer and founder of creative agency Istanbul 74, wrote prompts for Midjourney, a generative image AI platform. Integrating a faithful depiction of the handbags into the AI images required a hybrid approach. A product photographer shot the bags separately; a digital retouch artist then edited them into each scene, replacing AI-generated placeholders.
To Eşeli, it’s important that viewers understand the images are AI-driven rather than fully AI-generated. “AI is still not — and hopefully will not be — capable of creating without human commands. It depends on human ideas,” he says. “Like any form of creation, it starts with an idea. If you don’t have an idea, then AI won’t create one for you. AI is just a tool.”
There’s a degree of aesthetic overlap in all three campaigns. The models share the same uncanny poreless skin and promote the same ultra-thin beauty ideal. Their ethnicities are ambiguous but seem chosen with wide representation in mind. Backgrounds can appear suggested rather than defined; there’s a haziness that can make the imagery look plucked from a nearly forgotten dream. Hands, when visible, look not-quite human. Viewing an AI image can be an unnerving experience; everything is too smooth. “AI creates a nearly perfect image,” Eşeli says, “which can be great [but] sometimes can also be boring.”
Some audiences find generative AI campaigns chilling, perceiving in them an intent to reduce or eliminate the involvement of pesky humans. Last year, Levi’s faced backlash after saying it would use AI-generated models to increase “the number and diversity of our models”. In July, London-based SheerLuxe faced a similar backlash for introducing an AI-generated fashion and lifestyle editor. Called Reem Bot, it was met with widespread dismay by online commenters who criticised the platform’s leadership for creating an influencer of colour rather than adding to the diversity of the team by hiring an actual human.
Privately, many creatives and their representatives fret about the impact of AI. “I feel bad for the models,” one woman said at the Misela event. “Are they all going to be out of work?” Others maintain that there will always be a place for creativity of human origin.
“Regardless of technological advances, creatives will always be the driving forces that are setting the trends for the industries at large,” says Kent Belden, founder and chief executive of The Only Agency, which represents fashion stylists, make-up artists, photographers and other creatives. The hundreds of people involved in making a single music video can’t be replaced by a machine, he says. “All of those people have their own visions. Figuring out how to make those visions come together is not something that can be automated.”
Generative AI is an experimental exception rather than the norm in fashion — for now. “It’s progressing so rapidly,” Eşeli says. For their latest campaigns, both Etro and Misela returned to working with photographers; Mango has two more generative AI campaigns for its Teen line in the works. But with McKinsey forecasting that generative AI will add $150bn-$275bn to operating profits in the next three to five years, driven by marketing, it’s all but certain that there’s more to come.
“The best thing I learned from working on this campaign was the importance of the human touch in the process,” De Vincenzo says. “AI is just a new method to express creativity. Why not use it?”
Sign up for Fashion Matters, your weekly newsletter with the latest stories in style. Follow @financialtimesfashion on Instagram and subscribe to our podcast Life and Art wherever you listen
Read the full article here