Brands Using AI Models: Dated Cases and What Happened

How brands have used AI models so far

Brands have used AI models in three ways: to add range without new shoots, to make campaign images faster and cheaper, and to build licensed copies of real models. A fourth group has made a point of not using them at all.

We list only cases we could confirm in published reports; claims about other labels without a source we could open are left out. For how AI fashion models work, start with our main guide.

Read next: AI fashion models explained

Brands that used AI for speed and cost

Mango and Zalando are the clearest examples of AI models used to make more images, faster. Mango's first AI-generated campaign, in July 2024, was for its Mango Teen line. Mango said its own design, styling and photo teams made it, and its technology chief described AI as a co-pilot for the creative team.

Zalando went further. Reuters reported in May 2025 that about 70% of the editorial campaign images Zalando used in the fourth quarter of 2024 were AI-generated. The company said this cut production time from six to eight weeks down to three or four days, and costs by 90%. Zalando said traditional shoots are still needed. These are company figures, not independent measurements.

Licensed digital twins of real models

H&M is the main example of a brand building AI copies of real, paid models. In March 2025 it said it would create digital twins of 30 models. The models agreed to it, keep control and rights over their twin, and are paid for its use in a similar way to photo usage. H&M said early images would carry a watermark. Some influencers criticised the plan.

This route is now regulated in places. In New York, the Fashion Workers Act requires clear written consent before anyone creates or uses a model's digital replica. Our disclosure guide covers that law and the labelling rules.

Read next: AI model disclosure and consent rules

When AI models drew a backlash

AI models drew criticism when they seemed to stand in for real people, and a clear label did not prevent it. Levi's was the first big case: in March 2023 it announced a test of AI-generated models from Lalaland.ai to show more skin tones and body types. Critics said a brand should hire diverse models instead. Levi's then said AI models were not a means to advance diversity and that it was not scaling back live photo shoots or the use of live models.

In August 2025 a Guess ad in the US edition of Vogue showed models made by the AI agency Seraphinne Vallora. Critics online raised lost work for models and photographers, and unrealistic beauty standards. The agency's founders said their work is meant to supplement, not replace.

Labels did not settle it for luxury brands either. In December 2025 Valentino posted a labelled AI video for a handbag, and followers still called it cheap and lazy. In February 2026 Gucci posted labelled AI teaser images for Demna's first Gucci show, and comments called them AI slop.

When a generated face looked like a real person

A generated model can look like a real person, and that has already caused trouble. In September 2025 NBC News reported a Shein shirt listing priced at $11.69 whose model looked like a well-known real person; NBC could not confirm the image was AI-made. Shein said a third-party vendor listed it and it was pulled on discovery.

In June 2026 a New Zealand model said an image in a Huffer ad looked like a mix of him and his brother. The brand's managing director denied it and said no paid model's image had been manipulated. The claim was not settled in the reporting. Both cases show why a seller should look at the face in an AI image, not only the garment.

Brands that said no to AI models

Some brands now treat not using AI people as a selling point. Aerie, which pledged in 2014 not to retouch its models, shared a new pledge on Instagram in October 2025: no AI-generated people or bodies in its campaigns. Marketing Brew reported that the post became the brand's most popular, with more than 40,000 likes, and that Aerie's fourth-quarter 2025 sales were up 23% on a year earlier. The brand did not claim the pledge caused the rise.

Lessons for small sellers

The big cases point to a few habits for a small shop.

• Label AI people clearly. Labels did not stop criticism of luxury campaigns, but hiding AI use is worse, and in some places labelling is now a legal duty.
• Never use a real person's likeness without written consent. That covers copies of models you have worked with and faces that happen to look like someone famous.
• Check the face as well as the garment. If a generated model resembles a real person, make a new one.
• Do not use AI to fake range. A size or body type on a generated model does not mean the garment fits it. Only show sizes you sell, with real measurements.
• Keep real photos of the actual item. Most of the brands above kept live shoots alongside AI, and shoppers still need to see the real product.

Read next: AI fashion model prompts that keep the garment · Best AI fashion model generators · Clothes on a model without hiring one