Europe's biggest retailers are pushing back against the EU AI Act, and their argument hinges on a deceptively simple question: what counts as a deepfake?
According to The Decoder, Eurocommerce — the trade association whose members include Amazon, H&M, and IKEA — is asking regulators to exempt AI-generated advertising from the law's transparency requirements. Those rules are designed to force disclosure when content is artificially generated, the idea being that people deserve to know when an image or video isn't real.
Retailers say that logic doesn't fit how they actually use the technology. Their example, per The Decoder: an AI-generated picture of a living room used to sell a sofa. That image is synthetic, but it isn't a deepfake in the sense most people mean — it isn't impersonating a real person or faking a real event. Forcing a disclaimer on it, the industry argues, treats ordinary product marketing the same as deceptive synthetic media.
The scale of the issue is large. The Decoder reports that Zalando alone says 90 percent of the marketing content on its platform is already affected by the question.
The core problem, as The Decoder frames it, is that the EU hasn't drawn a clear line around what a deepfake actually is. Without a precise definition, a rule meant to fight deception risks sweeping in routine commercial imagery.
Why it matters: how the EU defines one word will decide whether millions of online product images carry "AI-generated" labels — and how much trust shoppers can place in what they see.