‘It’s unsettling’: AI agencies threaten to make fashion models obsolete

The next time you browse for Levi’s jeans online, you may be greeted by a gallery of computer-generated bodies and faces, all superimposed with the latest fashion.

Levi Strauss has become one of the latest fashion companies to incorporate artificial intelligence, aiming to “supplement” their roster of human models with hyperrealistic, AI-generated images later this year.

They’re not alone; with rapid advances in AI imaging technologies, entirely AI-based modelling agencies are popping up — and they’re only getting more realistic.

While AI is unlikely to replace well-known human models, experts say the tech is expected to hit many workers in the industry — especially entry- to mid-level models and support staff — extra hard. As some companies seek to use AI to emulate diversity, the question arises: Is any of this ethical?

Levi’s to feature AI in human clothing

Late last month, Levi Strauss announced they’d partnered with Amsterdam-based Lalaland.ai, a digital fashion studio that builds custom AI clothing models. Using generative AI, Lalaland specializes in creating hyperrealistic “avatars” of every “body type, age, size and skin tone” that can then be dressed in different outfits.

According to the studio’s website, it can take less than five minutes to generate an AI fashion model.

Lalaland’s website says it can generate a custom model in under five minutes. They charge 240 euros a month per user for up to 50 images, and 360 euros per user for unlimited renders.

Lalaland has also partnered with Calvin Klein and Tommy Hilfiger, according to Forbes, though the results of these collaborations are unclear.

In an email to the Star, a spokesperson for Levi Strauss said the company believes Lalaland’s technology could help them “publish more images of our products on a range of body types more quickly.”

In its original press release, Levi Strauss insinuated it would use the model’s customizability to inject more diversity into the fashion space. “We are not scaling back our plans for live photo shoots, the use of live models, or our commitment to working with diverse models,” the spokesperson added.

The announcement was immediately met with backlash online, with critics noting that AI images don’t represent real people and that the “diversity” generator may take the jobs of actual marginalized models down the line.

In response, Levi Strauss updated their release to read: “We do not see this pilot as a means to advance diversity or as a substitute for the real action that must be taken to deliver on our diversity, equity and inclusion goals and it should not have been portrayed as such.”

Lalaland.ai has not responded to the Star’s requests for comment before publication.

AI experts foresee significant job impacts

While certain models — especially ones with established brands — are likely safe from replacement, the same may not be true of many mid- to entry-level models, photographers, makeup artists and the myriad other support workers employed by the industry, said Richard Lachman, an associate professor at the RTA School of Media at Toronto Metropolitan University.

Richard Lachman, an associate professor at the RTA School of Media at Toronto Metropolitan University, said high-profile models are unlikely to be affected. He believes mid-to-entry level jobs will be the most impacted by AI.

“I think supermodels are going to be fine. The highest, most recognized characters that are appearing at the Met Gala are not really under attack from this kind of thing,” Lachman told the Star. “What’s in danger is the entry-level job or the mid-level job that gives someone regular paychecks. Those are (positions) that can be expensive for companies to have.

For reference, Lalaland charges 240 euros a month per user for up to 50 images, and 360 euros for unlimited renders. Other companies, like “virtual photo studio” Deep Agency, charge as little as $29 U.S. a month.

AI modelling also offers clients unparalleled customization, able to “match any set of specs that somebody wants,” Lachman said. The technology’s speed, cost and ease of use can seem “extremely attractive” to executives compared to hiring an entire team, spending hours shooting, flying people out to the location and more.

“Really, we’re starting to see this whole ecosystem endangered by these tools,” Lachman said. “ … People are harder to work with than software.”

Ishtiaque Ahmed, an assistant professor of computer science and a Schwartz Reisman Institute Fellow at the University of Toronto, told the Star BIPOC people and marginalized groups are more likely to be impacted by AI automation in general.

Generally speaking, BIPOC and marginalized groups are more likely to be impacted by AI automation simply because more of them work entry-to-mid-level jobs, said Ishtiaque Ahmed, an assistant professor of computer science and a Schwartz Reisman Institute Fellow at the University of Toronto.

“Historically, BIPOC people have been put in these kinds of jobs and that’s why they are in a higher risk,” Ahmed said. “ … If you compare an AI with a (disadvantaged person), you’ll see an AI is raised with a lot more privilege … it’s filled with a lot of education, it has a lot more power that a human being doesn’t get.

“So eventually, if you think about whether a company will get this robot or a human — they’ll definitely get the robot.”

Models and agencies speak out

Naomi Colford, a model signed with Toronto-based ICON Models Agency, told the Star she’s not too worried about her job right now.

“I can see why some people may think (AI) would be a beneficial and easy way to do (modelling),” she said, “but I think that it can never compare to having a true, authentic human as the model.”

Naomi Colford, a 22-year-old model signed with Toronto-based ICON Models Agency, isn’t that worried about her job for now. Models are more than their looks, she said; people gravitate toward their personalities, their brands. AI doesn’t have that yet.

AI will likely be cheaper than hiring a real model, she continued. “But when people, like myself, look at an advertisement, you feel more drawn to a real person — you’re more likely to buy that item if a real person is modelling it, I believe, over an AI.”

Colford conceded that AI images may eventually get good enough to become indistinguishable from photographs. At the moment, however, many AI models still look “off,” she said.

Even if the AI looked perfectly real, it can still never replicate a true human, Colford believes. Models, especially the high-profile ones, are more than their looks — people gravitate toward their personalities, their brands: “I just think as a society, we love to have certain people to look up to. And a lot of professional models are seen as role models, which with an AI, you will never be able to have,” she said.

At the same time, concern over AI is spreading throughout the industry, said Janelle Morgan, owner and director of Toronto-based Morgan Model Management.

Janelle Morgan, owner and director of Toronto-based Morgan Model Management, said brands’ use of AI to introduce diversity into fashion is “unsettling” and could lead to actual minorities being boxed out of the industry.

“We are concerned,” Morgan told the Star. “We are aware that this is happening, obviously, with brands now starting to reach out and booking AI models.”

“Right now, a lot of clients other than the really big ones are still used to just booking humans,” she said, but the future is uncertain. It can be tempting, especially for “designers with a shoestring budget,” to go with a far cheaper AI image rather than book a shoot through an agency, she said.

Digital diversity: Who do AI models represent?

As a Black woman who witnessed first hand the upstream battle to bring more diversity into modelling, Morgan said the “diverse AI sector” is “where I have a really big problem.”

“I think a lot of people don’t really truly understand the history when it comes to Black models and how we were boxed out,” she said. It hasn’t even been ten years since Naomi Campbell, Bethann Hardison and others fought for minorities’ right to the runway, she continued.

According to the Fashion Spot, a magazine monitoring diversity in fashion, 48.6 per cent of models surveyed in the fall of 2022 were people of colour. That’s a stark increase from seven years ago, when BIPOC models made up just 17 per cent.

“The doors have only been opened in the last like eight to nine years,” Morgan said. “ … So, it’s very jarring to know that brands are now saying, ‘Oh, we’re diverse,’ but they’re using AI and skipping over” working with actual minorities.

“It’s unsettling,” she said.

According to Lachman, AI being used to emulate diversity has troubling implications.

“The motivation (for fashion brands) was to create a set of models that look like society — a range of skin tones, body types, a range of ethnic backgrounds,” he said. “But in a sense, it’s producing an idealized, essentialist illusion of reality.

AI models “are not real people. They are not actually increasing representation, increasing the number of jobs,” he said.

“It doesn’t really reflect society — it produces a sort of fantasyland version of reality. And the effects of that are (concerning) in a very image conscious world.”

JOIN THE CONVERSATION

Conversations are opinions of our readers and are subject to the Code of Conduct. The Star
does not endorse these opinions.

For all the latest Technology News Click Here 

 For the latest news and updates, follow us on Google News

Read original article here

Denial of responsibility! TheDailyCheck is an automatic aggregator around the global media. All the content are available free on Internet. We have just arranged it in one platform for educational purpose only. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, all materials to their authors. If you are the owner of the content and do not want us to publish your materials on our website, please contact us by email – [email protected] The content will be deleted within 24 hours.