European companies have begun implementing required disclosures for AI-generated content as the European Union’s new transparency rules took effect on July 1, marking the start of enforcement under the EU’s landmark artificial intelligence regulation. The rules compel businesses to inform consumers when they are engaging with AI systems, including chatbots, deepfakes, and AI-generated images, audio, or video.

The regulations apply to a wide range of sectors, including online retail, travel, and advertising, prompting companies to update websites and customer service interfaces ahead of the deadline. The European Commission now holds the authority to impose fines on firms that fail to comply, with penalties reaching up to 3% of a company’s global annual revenue or €15 million ($17.3 million), whichever is greater.

Industry reactions have been mixed. Some companies, such as Berlin-based online retailer Zalando, which uses AI to create or enhance approximately 90% of its marketing content, criticized the Commission’s recently issued guidance as adding unnecessary complexity. Zalando expressed concerns that the lack of clear, consistent interpretation of the rules could lead to fragmented compliance across businesses.

Retail lobby group EuroCommerce cautioned against broad application of AI labelling requirements, warning such measures risk imposing excessive compliance costs on businesses while potentially confusing consumers. Similarly, CCLA Europe, representing major U.S. technology companies, argued that the guidelines’ broad definition of deepfakes could dilute the significance of AI labelling by requiring disclosures on almost all AI-generated content. Boniface de Champris, AI policy lead at CCLA, described these requirements as “burdensome” and likely to hinder innovation.

Conversely, EU officials and proponents of the regulations emphasize the importance of transparency in ensuring that AI content and interactions remain trustworthy. Henna Virkkunen, the European Commissioner for Innovation, Research, Culture, Education, and Youth, highlighted the need for clear disclosures given the widespread integration of AI in everyday life. Risto Uuk, head of EU policy at the Future of Life Institute, which advocates for stricter controls on high-risk AI, supported the rules as vital to informing the public about the capabilities and risks of AI technology.

Commissioners have pointed out that the guidelines include exemptions designed to prevent “labelling fatigue,” excluding certain uses such as standard photo editing. A Commission official stressed that the guidance was adjusted following industry consultations to ensure it is balanced and proportionate, avoiding transparency obligations that lack practical value.

Booking.com, an online travel service provider, welcomed the rules as aligning with growing consumer expectations. Pranav Pathak, director of AI product management at Booking.com, noted that being transparent about AI interactions can improve customer service outcomes. However, Pathak also called for clearer, simpler explanations from the Commission to help companies effectively implement the regulations.

The enforcement of these transparency measures is an initial phase of the broader AI Act, which came into force in August 2024. The legislation also grants the Commission powers to inspect general-purpose AI systems considered as posing systemic risks. However, requirements pertaining to high-risk AI applications have been postponed to December 2027 due to pressure from industry stakeholders and certain member states.

Experts advising businesses on the legislation describe the shifting regulatory landscape as a source of uncertainty, arguing that frequent changes and unclear implementation guidelines pose challenges for companies seeking to comply while continuing to innovate in AI technologies.