European retailers are pressing the EU to draw a clearer line between deceptive deepfakes and routine AI-generated advertising. The issue is becoming urgent for brands that already rely heavily on synthetic images for product marketing.
At the center of the dispute is a request from Eurocommerce, a European trade association whose members include Amazon, H&M, Inditex, and Ikea. The association wants AI-generated ad images that are not intended to deceive to be exempt from the EU AI Act's transparency requirements.
Why retailers want an exemption
In a letter to EU tech commissioner Henna Virkkunen, obtained by Reuters, Eurocommerce argued that some AI-generated advertising should not have to carry the same transparency labels as content that qualifies as a "deepfake." The EU law takes effect on August 2 and requires clear labeling of AI-generated or AI-altered content that falls under that category.
Eurocommerce's concern is practical as much as legal. Retail advertising now includes a large volume of images created or altered with AI. If the rule applies broadly, labels could appear on a major share of ordinary product marketing, even when the image is simply meant to present an item in context.
Director General Christel Delberghe offered a specific example: an AI-generated living room image used to showcase a sofa. Her argument is that this kind of image should not be treated like a deepfake, because it is not designed to mislead people about a person, event, or harmful scenario.
The association also argues that over-labeling could reduce the usefulness of transparency notices for consumers. If labels appear everywhere, including on low-risk product scenes, shoppers may find it harder to tell which warnings actually matter.
The sofa example shows the definitional problem
The core disagreement is not whether AI-generated content should ever be labeled. It is about what the word "deepfake" should cover when the content is commercial, non-deceptive, and focused on products.
A sofa shown inside an AI-made living room is different from an image built to impersonate someone or support fraud. Yet under the current framing described in the source article, both can fall under the same broad transparency discussion if the content is AI-generated or AI-altered and qualifies as a deepfake.
That creates uncertainty for retailers. A brand may use AI to build a room scene, adjust a campaign image, or create a model variation, but the business still has to understand whether the result triggers a disclosure duty.
The EU Commission has not responded to the demand yet. Until it does, retailers face a gray area: they know the law takes effect on August 2, but they do not yet have the narrower exemption Eurocommerce is asking for.
AI marketing is already part of retail operations
The debate is sharpened by how widely retailers are now using generative tools. Zalando says 90 percent of the marketing content on its platform is now AI-generated. That figure shows why even a small change in labeling rules could affect a large amount of day-to-day retail content.
Matthias Haase, VP of Content Solutions at Zalanda, described the operational shift this way: "Generative AI […] allowed us to move from a 'planning' mindset to a 'reacting' one, cutting those weeks of work down to just a few days — and our target this year is under 24 hours from spotting a trend to going live."
That quote explains why retailers care about the rule's scope. AI is not only being used to make isolated experiments. It is part of how some retail marketing teams respond to trends, prepare campaigns, and publish new content faster.
H&M and Zara also use AI-generated clones of models. That use case sits closer to the sensitive edge of the debate, because it involves representations of people rather than a product scene. Still, the source article presents both examples as part of a broader retail shift toward AI-generated marketing assets.
What is at stake for consumers
Transparency labels are meant to help consumers understand when they are looking at AI-generated or AI-altered material. Eurocommerce is not arguing against transparency in all cases. Its request is narrower: it wants an exemption for ad images that are not meant to deceive.
The concern is that a single label category may be asked to do too much. If it covers harmful impersonation, fraud-related content, model clones, and ordinary product staging, consumers may receive the same signal for very different kinds of content.
That matters because the term "deepfake" carries strong associations. The source article notes that the word has roots in non-consensual pornography and is mostly associated with fraud or other criminal activity. Applying the same term to an AI-generated sofa image can make the legal category feel broader than the everyday meaning consumers may expect.
The dispute therefore comes down to classification. Retailers want regulators to separate non-deceptive AI advertising from content that more clearly fits the deepfake label. Regulators, meanwhile, have not yet responded to Eurocommerce's request, leaving the industry waiting for clarity before the rules take effect.
A test case for AI rules in commerce
The Eurocommerce letter shows how difficult it can be to write AI rules that apply cleanly across many types of media. A transparency requirement can sound straightforward in principle, but retailers are asking how it should work when AI is used to create ordinary commercial images at scale.
The answer will matter for platforms, fashion brands, furniture sellers, and other companies that use generated visuals in advertising. If the EU treats a wide range of marketing images as deepfakes, businesses may need to label far more content than they believe the rule was meant to target.
If the EU grants the requested exemption, the key distinction would be whether an AI-generated ad image is intended to deceive. That would keep the focus on consumer protection while allowing common product presentation images to remain outside the deepfake category.
For now, the industry has a clear question and no public answer from the EU Commission. As August 2 approaches, the debate over AI-generated advertising is becoming a test of whether the EU AI Act can distinguish between synthetic media that misleads and synthetic media that simply sells.