Stop the fashion slop: Fake “AI” campaigns destroy creative work. Greenwashing makes it worse.

This September, United Colors of Benetton — the brand that once defined provocative fashion advertising under the creative direction of the late Oliviero Toscani — unveiled its Fall/Winter 2025 campaign. Dazed published a truly bizarre piece of sponsored content about the campaign, written as if it were the brand congratulating itself for its own work: “Benetton has shown that it’s possible to embrace cutting-edge technology while staying true to its legacy and the inclusive, pioneering spirit which has always driven it.” (You really need to squint to spot the #sponcon disclosure.)

Yes, Benetton presented what it paid Dazed to describe as a “diverse and smiling” cast, only… none of the models were real. Rather, this was a campaign prompted into existence by the Italian Instagrammer and synthetic media creator Rick Dick. Consequently, you can assume real models lost a potential gig to these kinds of synthetically generated faces, as did entire creative teams who'd normally produce such campaigns. 

Comments about the Dick campaign came fast and furious on the brand’s Instagram. “The wealthiest of brands are so greedy they’re willing to put an entire team out of work to save a few dollars,” one commenter wrote. “It’s not just models[;] it’s stylists, hair and makeup, assistants, photographers, editors, lighting [...] This will kill the careers of so many.”

Then there’s this gem about Dick: “‘Inclusivity’ campaign including every ethnicity, but the only paycheck goes to a white guy from Tuscany using AI. The audacity.” 

Of course, actual “diversity” would mean paying models, stylists, photographers. As one fashion photographer put it: “They be like ‘let’s prompt this into a generic stereotype based on thousands of pictures and blend them together into a fake model! That will save us so much work, money, and surely our clients couldn’t care less.’”

Perhaps the sharpest critique came from someone asking the obvious: “The impression of people wearing the impression of clothes? Can I use AI generated money to purchase?”

Model Denis Jovanovich summed it up: “We want real.”

Benetton is, of course, not the only purveyor of slop in the name of buzzword “diversity” and even greenwashed “sustainability.” 

This past winter, a photographer confided in me that one of their clients, a fashion brand, had found a novel way to reduce their carbon emissions — replacing product photography with “AI” generated slop. By no longer sending the photographer out to produce a campaign with clothes on human models, styled and produced by human creatives, the brand concluded it would no longer emit the small amount of emissions generated through its creative work (small, that is, when compared to the brand’s overall production). Of course, that also meant it'd conveniently not have to pay those humans, either — a microcosm of creative work in the mid-2020s.

It’s all part of a pattern we’ve seen accelerate over the past three years. In 2023, Levi’s was accused of “digital blackface” for using fake models to, ahem, “increase diversity,” without having to pay Black and brown talent, which led to a furious backlash. In February, Gucci commissioned Sybille de Saint Louvent to create a fully synthetic social media campaign, released the same day as the brand’s Fall/Winter 2025 show. De Saint Louvent specialises in “fake fashion campaigns,” according to the Business of Fashion, and she’s filled feeds with stylised slop released at a furious speed (“she initially set herself a challenge of producing one AI campaign a day to post to Instagram, though she has since relaxed her pace,” Marc Bain reported). The August issue of Vogue infamously ran an “AI” advertisement from Guess featuring a fake blonde white “woman,” a choice widely seen as a regressive promotion of unhealthy beauty standards. The slop goes on and on.

Tech companies are heavily promoting the use of “AI” imagery, often at a loss subsidised by massive bubble-like investment from venture capitalists, which floods the creative economy with cheap fake versions of things that human creatives could do themselves. Thus, it destabilises the market for working creatives in a way that’s analogous to the power differentials between fast fashion brands and suppliers.

One such startup, Photoroom, has heavily promoted itself to consumer brands across industries including fashion and beauty, with the promise that a poorly lit iPhone photo of, say, a flatlay dress can then be synthetically staged on a fake model. Or a full face of makeup can be synthetically generated without the need to pay a makeup artist.

More cynically, Photoroom has tried to lure ethically minded users with claims that “AI” image generation could “cut [a brand’s] carbon footprint.” The photographer confiding in me — who requested to remain anonymous for fear of losing out on work — received a link to such an article published by Photoroom in 2024, which makes this dubious case.

The Photoroom blog reads: “A recent study reveals that the carbon dioxide equivalent emitted from creating an image with AI is significantly lower, ranging from 310 to 2900 times less, than the emissions associated with human-led photo production.” 

Continuing, the Photoroom blog writer glibly argued: “We’ve identified an impactful benefit for our customers: beyond selling more, there are environmental benefits of using digital editing over traditional photoshoots” — failing to address the contradictions inherent in this one greenwashed sentence.

This post cites a strange 2023 study, which has myriad problems, beginning with the rather murderous implication that we could “replace” human workers to “reduce” a brand’s “carbon footprint.” Indeed, “carbon footprint” was itself a PR distraction created by fossil fuel company BP with its advertising agency, so it’s not all that surprising that tech companies selling environmentally harmful, energy guzzling technologies might borrow that framing to promote themselves.

That study’s methodology was subsequently skewered on an episode of the critical tech podcast Mystery AI Hype Theater 3000, hosted by two of the leading critics of Big Tech’s unprofitable race to develop “AI”: University of Washington professor Emily M. Bender and the Distributed AI Research Institute’s Alex Hanna, a former senior research scientist at Google.

In their new book The AI Con, which launched in the UK this May, Bender and Hanna dissect the hype around these technologies. Their message is not doomerism but “a sense of renewed possibility,” Bender told me in the lead up to the book’s release. “Both in terms of what can be achieved through solidarity and in terms of arguments that can be presented to those who buy synthetic media.”

In fact, the authors directly address the anxieties that creative workers like my photographer friend — and those of us who value human creativity more broadly — feel in this moment. “We wrote this book with a lot of creative folks in mind,” Hanna told me in May. “Many of them are not unionised and support themselves doing freelancing, which makes them particularly precarious. One of the things which we find really heartening is how people lean into their expertise and focus on their craft in the face of so much AI slop and job displacement.”

What then is the real motivating factor for all this fashion slop: carbon emissions or the avoidance of paying creative workers?

Certainly, there are emissions related to creative work that brands should be calculating. Photography and campaign production fall within Scope 3 emissions, which encompasses all indirect emissions in a company’s value chain. Given that Scope 3 often constitutes the vast majority of a fashion company’s total emissions, accounting for marketing-related activities is a good thing to do. But photographers and make-up artists are not driving the majority of a brand’s emissions.

If a brand really wanted to reduce its emissions, creative work is not the first place to cut. Decarbonising manufacturing is the core area where brands can achieve immediate and practical reductions right now, without the need for new technology, as outlined in the 2025 “What Fuels Fashion” report from Fashion Revolution. 

Yet if you look at data about the fashion industry’s environmental performance gap, such as this 2025 Good On You report by journalist Megan Doyle, it’s clear that most brands are doing very little of that actual work. Many set targets and then fail to disclose any progress against them. Whether greenwashing or greenhushing, the underlying system plugs along, only now with “AI!”

Creative workers, many of whom remain freelance and contract-based, are often hesitant to speak publicly because speaking up means being seen as difficult. They’re passed over for other work. Even the reduced budget in this economy can seem better than nothing. (Once I was on a date with a celebrity makeup artist in London, who confided that he had lost fashion and beauty work as a client had uploaded images of his prior campaign work to a chatbot to generate synthetic images bearing his signature style, but without any compensation.)

Indeed, this is one of many examples of how “AI” — by which we mean synthetic media extruding machines, to borrow the verbiage used by Bender and Hanna in The AI Con — is making work in the creative industries worse. Flooding the creative market with this austerity technology is already making fashion sloppier.

But resistance is an option. As Bender and Hanna argue in their book, we can reject inevitability narratives; we can make slop taboo; and, at a minimum, we can counter the nonsense hype claims that replacing human creatives with “AI” is a climate solution — as startups like Photoroom absurdly promote.

Where do we begin? One way is to think about “AI” much in the same way that we think about fast fashion, an analogy the fashion brands eagerly adopting the tech would rather ignore but which the leading accountability journalists in the “AI” space make themselves.

Earlier this year, I interviewed investigative journalist Karen Hao, author of the instant best-selling book Empire of AI, which traces the Big Tech development of “AI” as we know it today through the framework of empires of old.

Hao was the first journalist to profile OpenAI, the creators of not only ChatGPT but the APIs that many startups essentially white label. I was intrigued when she told me that early on in her career as an investigative journalist covering “AI”, she looked to the accountability reporting in another industry — fashion — to inform her reporting in tech.

“I sort of analogise it a little bit to fair trade and the fashion supply chain,” Hao told me. “There was such great reporting at the time on the fashion industry and how we do not actually need clothes that lead to human rights abuses.”

This struck me as ironic that many people who work in so-called “sustainable fashion” did not seem to be making that connection themselves or applying the same kind of value chain scrutiny to generative “AI” systems that they apply to fashion. In our interview, I relayed to Hao that many people working on labour and environmental justice issues in fashion had seemed to have drunk the Kool-Aid.

She told me the media has a role to play in helping the public see that “AI doesn’t just arrive fully formed […] There are real economies of scale, economies of extraction, labour exploitation and data exploitation that are churning beneath the surface to produce this technology in the first place.”

To think through whether the use of “AI” in fashion is “ethical” or “sustainable,” then, requires a shift from imagining “AI” as something in the “cloud” and instead to understand it for what it is: a very physical and tangible product. 

The current mode of “AI” development harms people and the planet in many ways. The energy demands have driven Big Tech’s indirect emissions up 150% from 2020 to 2023, a recent United Nations report concluded. The surge in data centres runs fresh water dry in affected communities, as Hao has reported. It’s a labour justice issue, too. As with the production of our clothes, “AI” requires a vast but intentionally obscured workforce in the majority world — creative and talented individuals who suffer post-traumatic stress disorder and other mental health challenges due to their experiences working jobs ranging from content moderation to data labelling, as DAIR has investigated in its landmark Data Workers’ Inquiry. And like ultra fast fashion steals from independent designers, “AI” systems are trained on the brazenly stolen works of artists, writers, fashion designers, photographers, make-up artists, models, and many more creatives who have not consented to or been fairly compensated for their labour. 

“AI” then is to creative work what ultra fast fashion is to fashion. And as you can see in the comments beneath many such campaigns, some people are resisting the slop in the same way they resist Shein and Temu’s race to the bottom.

Do synthetic media generators like Photoroom or ChatGPT emit more or less emissions than a human creative team? Is Puma’s experimental “AI” campaign or the annual AI Fashion Week more or less sustainable than human created fashion campaigns? Is replacing human creatives with “AI” better for the planet? These are the wrong questions to be asking.

Replacing humans with “AI” in the name of “sustainability” is eco-colonialism, white supremacist, and deeply misguided. The only way to meaningfully reduce fashion’s carbon emissions is to actually reduce the amount of carbon emitted in making more clothes every year than the world's population could ever wear.

No image generators are going to make a dent in fashion’s take-make-waste system.

Rather, the tech companies making those carbon footprint claims about replacing creatives with synthetic media generators are simultaneously promising that “AI” will help brands sell more clothes, faster than ever before. They’re advertising their tech as a way to both reduce and increase emissions, and that’s untenable.

As Bender and Hanna write in their book: 

I admit: With so many overwhelming problems in fashion, from forced labour to poverty wages, “AI” disrupting the creative economy can seem like yet another overwhelming thing to worry about. 

There is, however, one thing I’m clear on: If a garment is designed and created by human hands to last for a long time, then it should be modelled by humans. If a brand’s campaign only shows fake clothes on fake people, then there’s a good chance their clothes are only fit for fake people and only worth Monopoly money. Or, at least, that’s the editorial policy for anxiety.eco.

The things that I still love about style, design and life all stem from creativity — the real kind; the kind the Big Tech companies stole to create their slop generators; the kind that can never be replaced with their uncanny slop.

Because slop is anything but fashionable, darling.