Browse otto.de’s womenswear section and the models look right. Good posture, plausible skin texture, fabric that sits the way a polyester blend sits on a size-38 frame. They are synthetic, each generated in minutes by Otto’s MOVEX Virtual Content Creator from a single product photograph — and no label was visible on the product detail pages reviewed for this piece.

The economics are plain. Otto reports up to 60 percent lower production costs and five times the daily content output of traditional shoots. It is already offering the platform to other retailers. Zalando, per an analysis in Chief AI Officer, generates roughly 70 percent of its editorial images with AI at 90 percent lower campaign costs. H&M photographs real models, then builds “digital twins” for social campaigns.

Three of Europe’s largest fashion platforms have converged on the same cost logic within eighteen months, and none has established a consumer-facing disclosure standard.

H&M released a behind-the-scenes film alongside its digital twins campaign and calls the approach “transparent”; what that looks like on a product detail page is not part of its public record. Otto’s press materials celebrate the technology’s speed, diversity, and cost efficiency without addressing consumer-facing disclosure. A Chief AI Officer analysis of Zalando’s AI rollout centres on quality review; consumer transparency is not among the priorities it describes. The gap shows in demand data: a Stylitics survey of 411 shoppers found 59 percent want clear labeling of AI product images, yet in a side-by-side comparison, 71 percent said they looked the same as or had only small differences from photographs.

The EU AI Act’s Article 50 should close part of this gap. Beginning August 2026, providers of AI systems generating synthetic images must ensure outputs are marked in a machine-readable format. The European Commission published a first draft Code of Practice for AI labeling in December 2025; the final version is due by June 2026. The Draft Code requires deployers to clearly label AI-generated content at the point of first customer interaction, but it specifies no format for how that obligation applies to individual product images — leaving retail PDPs in a grey zone until the final text arrives.

Adel Rootstein’s hyper-realistic mannequins arrived in London shop windows in the 1960s, sculpted from living models, Twiggy among them. Retailers gained synthetic likenesses of real people, posed and lit to sell product, with no obligation to name the distance between the figure in the glass and the body that inspired it. Sixty years later the synthetic likeness is digital, the display window is a product detail page, and the disclosure obligation is still being drafted.

Otto is not doing anything unusual. It is doing what the cost curve demands, inside a regulatory vacuum that rewards the first mover. The standard the industry needs is a visible, consistent disclosure norm that treats the shopper’s right to know what they are looking at as a design requirement — and nobody has written it yet.