Retailers seek AI-ad exemption from EU disclosure rules
TL;DR:
- Eurocommerce, the European retail body, has asked the EU to exempt AI-generated advertisements from new transparency rules.
- The EU AI Act, in force from 2 August, requires labelling of AI-generated “deep fake” images, video or audio.
- Retailers including Amazon, H&M and Inditex argue benign product imagery should not count.
A lobbying skirmish has opened over how the EU AI Act’s transparency obligations apply to advertising. Eurocommerce — whose members include Amazon, H&M, Inditex and Ikea — has written to the EU’s tech chief, Henna Virkkunen, urging that AI-generated ads be carved out of disclosure rules that take effect on 2 August.
What counts as a “deep fake”?
The Act requires companies to label content where AI has generated or modified images, video or audio “constituting a deep fake”. Eurocommerce director general Christel Delberghe argues that ads not intended to mislead — her examples include generating a living-room scene to showcase a sofa, or enhancing product visuals — should fall outside that definition. The trade body’s concern is scale: retailers already use AI to produce marketing imagery widely, with Zalando saying the technology cut content-production costs by 90% and fast-fashion firms deploying AI-generated model clones. Labelling “a very large share” of that output, the letter warns, would dilute the value of disclosure to consumers. The European Commission has not yet commented.
The dispute matters beyond the EU’s borders. UK retailers selling into the single market will be bound by these rules regardless of domestic policy, and Britain’s own approach to AI transparency in advertising is still taking shape. The Advertising Standards Authority has already warned advertisers about AI content and deepfakes, suggesting the labelling question is live on both sides of the Channel.
Looking forward
The case captures a recurring tension in AI regulation: rules written to catch deceptive synthetic media inevitably sweep in mundane commercial uses, forcing regulators to draw lines between manipulation and mere efficiency. How Brussels responds will set an early precedent for how broadly “deep fake” disclosure bites — and UK firms, watching for read-across, have good reason to track where that line lands.