Why Fashion Doesn’t Talk About How It Uses AI

Subscribe to Tech Mode by Marc Bain to go deep on the most intriguing developments in AI and how technology is reshaping the fashion industry.

Welcome back to Tech Mode, your monthly guide to how AI and other technologies are reshaping the fashion industry.

One of the things that interests me about the large language models shaking up the world is that some of their potentially impactful uses aren’t immediately obvious. Sure, writing copy and generating ad imagery were no-brainers for fashion companies. But when Matt Kropp, CTO of BCG X — an arm of the consultancy focused on designing and building AI solutions for clients — told me last year about how BCG was using LLMs to generate market research, giving companies insight into what shoppers are thinking, the idea surprised me. I had never thought about businesses deploying LLMs that way.

It turns out a cottage industry has sprung up around synthetic consumer research, where AI is used to simulate shoppers and provide brands with information they can apply in developing products, marketing campaigns, discount strategies and more. From my conversations with people like Kropp and Kelly Pedersen, a partner at PwC who leads the consultancy’s global retail practice, it’s clear a number of big companies are already using synthetic research — even if they aren’t talking about it publicly yet. Target is so far the only large retailer I’ve noticed plugging its use of synthetic audiences.

The idea has its strengths and weaknesses, which you can read about in this story we published last week, but it’s always intriguing to me how cautious big fashion and retail players are about disclosing the ways they’re using AI.

On that note, let’s dive in.

Why Retailers Are Quiet About AI

Widespread usage: UBS analyst Jay Sole actually tackled the question of how many companies in his coverage are using AI in a note to investors Tuesday. He based his analysis on disclosures across things like press releases and company filings, earnings calls, investor day conferences and conversations with company management. The answer was 43 out of 45 companies are using AI, which is to say nearly all of them. (Birkenstock and Dillard’s were the only two companies he found with no specific examples of AI adoption.)

From that data, you’d think companies were constantly blabbing about how they’re employing the technology, yet Sole noted that, in his discussions with companies public and private, there was a good deal of sensitivity around disclosing AI use. “Our view is the reason is companies likely don’t want to give away potential sources of competitive advantage,” he wrote.

Leaders and laggards: Sole also created an AI scorecard ranking companies on their use of the technology across six different dimensions: product development and innovation, supply chain and logistics, marketing, sales and customer experience, store operations and, lastly, organisation and support functions. Of the 45 companies, 14 scored 100 percent, meaning they had some level of AI adoption in every dimension, with those attaining a perfect score including retailers from Revolve and Levi’s to Steve Madden and Victoria’s Secret.

Across all the companies examined, use of AI was most prevalent in sales & customer experience, though Sole noted that adoption of the technology varied widely, sometimes even within a company. American Eagle uses AI for back-office process efficiency, but its Aerie brand has also pledged never to use AI-generated models, for example.

Let’s talk about it: The reticence about AI in fashion at first seems unusual considering how companies in some other industries talk about the subject. The phrase “AI washing” has even emerged because of how many businesses claim to be using AI when they’re not. And it’s not exactly a shameful thing to hide. Sole noted that AI can make retailers more efficient. “In fact, we believe companies like [Ralph Lauren] and [Tapestry] have been able to deliver very strong results recently in part due to AI,” he wrote.

It makes more sense, however, when thinking about the challenges of implementing AI. A company may be reluctant to talk about a pilot project it’s still not sure will work out, or a tool that may be impactful for a small group of employees but might not be moving the needle at the company level, or a change in the workflow that holds a good deal of promise for the future but hasn’t yet yielded measurable gains. The rise of publicly available LLMs also means plenty of employees at companies are using AI informally. These are all uses companies might not want to go shouting about.

Of course, others could be what Sole said: competitive advantages companies don’t want to share with rivals.

“Addictive Design” on Trial

Line of attack: Regulators and plaintiffs in lawsuits have a new favourite way of going after social media companies for alleged harms, and the strategy has now reached fashion. Last week, the EU opened an investigation into Shein over concerns that included Shein’s “addictive design.” The probe will consider whether features of the platform like awarding points for engagement could be harmful to users’ wellbeing, and the transparency of its recommendation system will also be under scrutiny.

Addictive design has become a prominent new line of attack against social networks. The EU previously warned TikTok that it may be in breach of the bloc’s Digital Services Act because of features like infinite scroll that keep users hooked on the app. Last week, a major trial also began in the US stemming from a lawsuit accusing Instagram and YouTube of causing mental harm with their addictive designs. (Snap and TikTok already settled with that plaintiff.)

What’s at stake: The common thread in these cases is the allegation that the platform built an addictive experience that can be harmful to users’ mental health, and the outcomes could be big alterations to those experiences. The EU told TikTok in a preliminary decision that it will need to change its design, phasing out features like the infinite scroll and implementing others like “screen time breaks.” If TikTok doesn’t come up with solutions, it will be hit with an order to re-engineer its app and a fine of up to 6 percent of total worldwide annual turnover.

In the US, the trial against Instagram and YouTube is the first of more than a thousand such cases lined up against the big social networks. If they lose, it opens the door to massive claims of damages, and could also force a rethink of the platforms’ designs. (Incidentally, Instagram’s beauty filters and their effect on teen girls’ mental health were a hot topic in the courtroom as the plaintiff’s attorney grilled Meta founder and chief executive Mark Zuckerberg last week.)

What it means for fashion: For most retailers aside from Shein, these actions may not have a direct and drastic impact. Shein is under scrutiny in the EU because, under the Digital Services Act, it qualifies as a “very large online platform,” meaning it has more than 45 million users in the EU. Few fashion retailers are as large, and even fewer have apps with reward systems like Shein’s.

But given that these platforms are some of fashion’s favourite marketing channels, any changes they’re forced to make could still have consequences for brands and retailers. Right now, fashion businesses benefit from the way social media algorithms surface fashion content to the users who are most likely to engage with it, whether that’s campaign images or posts by influencers peddling products through affiliate links. Those benefits aren’t likely to suddenly disappear, but they could be moderated, or brands could have to adjust strategies to new environments where users are suddenly discouraged from spending countless hours scrolling.

Brands Rethink Investment in Virtual Worlds

Pressing pause: Brands significantly pulled back their investments in virtual worlds like Roblox, Fortnite and Zepeto last year, according to a new report by Geeiq, a gaming data and strategy firm. After rocketing from $3 million in 2020 to $441 million in 2024, investments plunged nearly 50 percent in 2025 to a more modest $229 million.

The total number of new brand activations also declined last year, dropping 17 percent.

Game on: That doesn’t necessarily mean brands lost all interest in gaming, however. While they pulled back on creating fully owned, standalone destinations within virtual worlds, such as E.l.f. Beauty’s My Makeup Store on Roblox, they launched more integrations, which are more like branded activations inside existing games. An example would be Valentino working its fragrance into the game Murdery Mystery on Fortnite.

These types of integrations were up 14 percent, even as the number of new brand-owned worlds shrank by 57 percent.

What’s the deal: Geeiq’s take on the data was that it marked an inflection point where brands are moving from experimentation to “more selective, goal-led activation.” They’re doing fewer big, one-off launches and instead shifting to more carefully planned activations that are quick to create, easier to measure and offer repeatable results.

The shift sounds like what happens when brands reach a certain level of maturity in what was once a new channel. In the early 2020s, labels like Balenciaga, Gucci and Vans were rushing into the world of gaming to plant their flags. In 2026, being present in Roblox or Fortnite is no longer a novelty, while companies are undoubtedly looking back to see if the investments paid off. Not all of them did. The question is what comes next, and how do brands make it worthwhile.

Want to dive deeper into an insight from this article? Check out The Brain of Fashion, BoF’s new generative AI tool where you can unlock BoF’s technology archive with a single question.