Blab AI
Retail Association Urges EU to Exempt AI-Generated Ads from Deep-Fake Disclosure Rules
Eurocommerce, the European retail association that counts Amazon, H&M, Inditex and Ikea among its members, has asked the European Commission’s technology chief, Henna Virkkunen, to exclude AI‑generated advertisements from the EU’s new transparency rules.
In a letter sent to Virkkunen on Thursday and reported by Reuters on Friday, Eurocommerce’s director‑general Christel Delberghe argued that the EU Artificial Intelligence Act’s requirement to label content that constitutes a deep fake should not apply to AI‑generated ads that are not intended to mislead consumers. The association said that such ads, for example, an AI‑generated image of a living room used to showcase a sofa or an enhanced product photo, should not be treated as deep fakes.
The EU AI Act, which entered into force on 2 August 2024, requires companies to clearly label any image, video or audio that has been created or modified by artificial intelligence and that falls under the definition of a deep fake. The regulation is part of the EU’s first comprehensive AI framework, which classifies AI systems by risk level—unacceptable, high, limited, minimal—and imposes transparency obligations on limited‑risk applications.
Eurocommerce warned that applying the deep‑fake labeling requirement to AI‑edited or AI‑generated advertisements would compel retailers to mark a large portion of their marketing content. The association said that such widespread labeling would dilute the usefulness of the disclosure for consumers.
Retailers already use generative AI extensively for marketing. German online retailer Zalando has reported that the technology has cut its content‑production costs by 90 percent and reduced the time needed to produce imagery from several weeks to a few days. Fast‑fashion brands such as H&M and Zara have employed AI‑generated clones of models for advertising.
The letter to Virkkunen did not include a response from the European Commission. The Commission has not yet issued a statement on the request.
The EU AI Act’s transparency obligations for deep fakes are designed to protect consumers from deceptive or harmful synthetic media. The Act’s definition of a deep fake focuses on content that has been manipulated to create a false representation of reality. Eurocommerce’s position is that purely illustrative or enhancement‑type AI content does not meet that definition.
The request comes amid ongoing debate about the scope of the AI Act’s disclosure rules. Some industry groups argue that the regulations are too broad and could stifle legitimate uses of generative AI in advertising, while regulators emphasize the need for consumer protection.
As the EU AI Act is implemented gradually over the next several years, the outcome of Eurocommerce’s appeal could influence how the transparency requirements are applied to the retail sector. The association’s letter highlights the tension between the benefits of AI‑driven marketing efficiencies and the regulatory framework’s intent to ensure that consumers are informed about synthetic content.
At present, the EU has not announced any changes to the AI Act’s deep‑fake labeling provisions. The European Commission’s response to Eurocommerce’s request remains pending.
The debate is likely to continue as more retailers adopt generative AI tools and as the EU finalizes its enforcement strategy for the AI Act. The outcome will determine whether AI‑generated advertisements will be required to carry a disclosure label or whether they will be treated like conventional marketing content.
For now, the EU AI Act remains in force, and retailers must comply with its transparency obligations for any content that falls within the deep‑fake definition. The industry’s request for an exemption is a reminder that the practical application of the regulation is still being negotiated.
The situation underscores the need for clear guidance on what constitutes a deep fake in the context of advertising and how the EU’s AI Act will be applied to the rapidly evolving landscape of generative AI marketing.
In a letter sent to Virkkunen on Thursday and reported by Reuters on Friday, Eurocommerce’s director‑general Christel Delberghe argued that the EU Artificial Intelligence Act’s requirement to label content that constitutes a deep fake should not apply to AI‑generated ads that are not intended to mislead consumers. The association said that such ads, for example, an AI‑generated image of a living room used to showcase a sofa or an enhanced product photo, should not be treated as deep fakes.
The EU AI Act, which entered into force on 2 August 2024, requires companies to clearly label any image, video or audio that has been created or modified by artificial intelligence and that falls under the definition of a deep fake. The regulation is part of the EU’s first comprehensive AI framework, which classifies AI systems by risk level—unacceptable, high, limited, minimal—and imposes transparency obligations on limited‑risk applications.
Eurocommerce warned that applying the deep‑fake labeling requirement to AI‑edited or AI‑generated advertisements would compel retailers to mark a large portion of their marketing content. The association said that such widespread labeling would dilute the usefulness of the disclosure for consumers.
Retailers already use generative AI extensively for marketing. German online retailer Zalando has reported that the technology has cut its content‑production costs by 90 percent and reduced the time needed to produce imagery from several weeks to a few days. Fast‑fashion brands such as H&M and Zara have employed AI‑generated clones of models for advertising.
The letter to Virkkunen did not include a response from the European Commission. The Commission has not yet issued a statement on the request.
The EU AI Act’s transparency obligations for deep fakes are designed to protect consumers from deceptive or harmful synthetic media. The Act’s definition of a deep fake focuses on content that has been manipulated to create a false representation of reality. Eurocommerce’s position is that purely illustrative or enhancement‑type AI content does not meet that definition.
The request comes amid ongoing debate about the scope of the AI Act’s disclosure rules. Some industry groups argue that the regulations are too broad and could stifle legitimate uses of generative AI in advertising, while regulators emphasize the need for consumer protection.
As the EU AI Act is implemented gradually over the next several years, the outcome of Eurocommerce’s appeal could influence how the transparency requirements are applied to the retail sector. The association’s letter highlights the tension between the benefits of AI‑driven marketing efficiencies and the regulatory framework’s intent to ensure that consumers are informed about synthetic content.
At present, the EU has not announced any changes to the AI Act’s deep‑fake labeling provisions. The European Commission’s response to Eurocommerce’s request remains pending.
The debate is likely to continue as more retailers adopt generative AI tools and as the EU finalizes its enforcement strategy for the AI Act. The outcome will determine whether AI‑generated advertisements will be required to carry a disclosure label or whether they will be treated like conventional marketing content.
For now, the EU AI Act remains in force, and retailers must comply with its transparency obligations for any content that falls within the deep‑fake definition. The industry’s request for an exemption is a reminder that the practical application of the regulation is still being negotiated.
The situation underscores the need for clear guidance on what constitutes a deep fake in the context of advertising and how the EU’s AI Act will be applied to the rapidly evolving landscape of generative AI marketing.
