This article on digital celebrity is part of our Vogue Business membership package. To enjoy unlimited access to our weekly Technology Edit, which contains Member-only reporting and analysis and our NFT Tracker, sign up for membership here.
In April, British singer-songwriter FKA Twigs informed American senators on an intellectual property Senate subcommittee that she used artificial intelligence to develop a deepfake version of herself. She trained her digital replica to imitate her personality and voice. The AI Twigs is expected to appear on social media later this year, freeing up the artist to focus on her art.
FKA Twigs was entirely comfortable with her creation. “As a future-facing artist, new technologies are an exciting tool that can be used to express deeper emotions, create fantasy worlds and touch the hearts of many people,” she said in a statement.
More practically, she’s determined to take control of AI for herself rather than allow unauthorised AI-generated deepfakes to exploit her image. “These and similar emerging technologies are highly valuable tools both artistically and commercially when under the control of the artist,” she said.
Following in her footsteps, other celebrities are now partnering with technology companies to develop digital twins with generative AI that can engage with fans on their behalf. Soul Machines, a New Zealand-based company, has offered 24-7 chats with digital celebrities including K-pop singer Mark Tuan, MMA fighter Francis Ngannou and retired golfer Jack Nicklaus. Musician Grimes, who is a strong proponent of technological experimentation, last year unveiled software that invites anyone to make AI-generated songs using her voice.
In September, Meta launched AI chatbots played by celebrities such as Kendall Jenner and Tom Brady, but so far they can only interact with users on Meta’s social media apps. TikTok recently launched a product that will allow brands to create generative AI avatars based on creators and paid actors, as well as an AI dubbing tool that lets creators translate their content into 10 different languages.
Digital twins can save time
Even before the emergence of generative AI, the fashion industry was exploring the marketing potential of virtual humans. In 2022, Puma released a lifelike 3D avatar of football star Neymar Júnior during New York Fashion Week. High-profile veteran model Eva Herzigová announced her digital twin last year — and it can even be hired by brands.
Virtual humans are, in some ways, easier to work with. “The way in which those campaigns are executed relies on specific artistic direction or stylisation or effects, which are more difficult to achieve if you are working with the physical talent,” says Simon Windsor, co-CEO of Dimension Studio, the company behind the creation of Neymar and Herzigová’s digital twins. “It’s not the case that these virtual models are being used to replace a traditional brand campaign; it’s just that they are opening up new possibilities.”
Working with a digital human can be time efficient because there is no need for all the complexities of a physical photoshoot. Brands also have more control over the creative process, says Rahul Titus, global head of influence at advertising firm Ogilvy. “Do you really want a big celebrity to sit there, shooting 500 variations of an ad, when I can licence that?” he says. While the campaign might still require approval from the celebrity, Titus says the technology “unlocks a whole world of potential”.
Dimension Studios has created virtual humans for Balenciaga, H&M, Nike and Charlotte Tilbury. Fashion and beauty brands have also turned their virtual personas into digital spokespeople: LVMH created a virtual ambassador called Livi, while Nars Cosmetics revealed three meta-humans. Another AI startup, Metaphysic, created an ad campaign in 2021 for Gillette that featured former American football player Deion Sanders, depicted in his 1989 prime.
Few fashion and beauty brands have established policies or guardrails around such uses, with some showing resistance. Alexandra Bolten-Mercadier, global head of social and content marketing at L’Oréal, said recently that the brand will “not use generative lifelike face, body or skin” in external communications.
Creator agency Billion Dollar Boy advises against using generative AI to create or recreate a human’s likeliness. However, the agency has seen more engagement with AI-generated imageries and videos that are creative and outlandish; a Versace campaign made in partnership with Billion Dollar Boy, for example, showcased the maison’s Greca Goddess handbag in a new way and received a 1,460 per cent higher play rate.
“Consumers [are] very confused,” Titus says. “People trust people more than they trust brands, which is why we use these celebrities in the first place, and when suddenly you are starting to use AI, that authenticity is completely lost.” Last year, Ogilvy launched an AI Accountability Act, which included a call for brands to disclose AI use cases.
Virtual models have already become a source of controversy. Back in 2018, dark-skinned digital model Shudu was criticised for being the creation of a white man and taking away opportunities from real Black models, while last year, Levi’s use of AI-generated models drew backlashes for similar reasons, as critics saw it as a cheap way to address the issues of representation. In May, Scarlett Johansson clashed with ChatGPT parent OpenAI over a voice assistant that sounded just like her after the star declined permission to use her actual voice — a cautionary tale for brands who are entering the space. As people increasingly become aware of unauthorised uses of AI, social media platforms like TikTok, Instagram, Facebook and YouTube all require labels for AI-generated content.
Data security concerns
Legal experts say brands should prioritise establishing data security protocols. When a talent or agency allows a brand to borrow a digital twin, their data, including body measurements and facial specs, could be lost or hacked, says entertainment lawyer Mikey Glazer. A potential solution is the use of blockchain, a digital ledger associated with an asset (such as the data for a digital twin). If the data is hijacked, the digital twin could be deactivated.
The tech could help to protect talent, especially those with smaller platforms, via allowing them to take further control of their identities, says Jade McSorley, a former model and PhD researcher focusing on technology’s impact on fashion. Since transactions of any asset are recorded on the blockchain, the talent knows when digital twins are used in ways “outside the bounds of the contract”, and has evidence to take action.
That’s especially important for young models who don’t often have a say when their agencies create digital twins based on their likeness, she adds. “Those models and actors who are recognisable have a lot more rights than working fashion models, i.e. myself when I was modelling,” McSorley says. “You are not recognisable, so there is actually not as much we can stand on if our digital version gets misused in any way.”
Becky Owen, CMO of Billion Dollar Boy, warns brands and talent to proceed with care, as the risks of using digital twins go beyond data breach. As technology advances to allow digital clones for everyone, Owen says that it could also perpetuate the stereotypes that fashion has long tried to shy away from.
“If you are getting a body double, are you actually airbrushing yourself? Are you creating an imperfectly perfect version of yourself?” she says. “It won’t have any of your stretch marks or dry skin patches. That is pushing us into a world that we don’t want to be.”
Clarification: This story has been changed to clarify a sentiment around the ease of working with virtual celebrities. (2 July 2024)
Sign up to receive the Vogue Business newsletter for the latest luxury news and insights, plus exclusive membership discounts.
Comments, questions or feedback? Email us at feedback@voguebusiness.com.
