EU Expands Transparency Rules for AI Media

The EU just rolled out a transparency push for AI media. The move ties a voluntary EU Code of Practice to the broader task of implementing the EU AI Act, inviting providers and deployers of generative AI to disclose how content was created and edited so audiences can judge authenticity.

The release, published May 6, 2026, comes as AI tools become harder to distinguish from human-created material. The Partnership on AI explains that the Code of Practice is being shaped through a multi-year process that aims to align industry behavior with regulatory aims, while the EU’s own lawmaking progresses in parallel. A new advisory body, the SAIGE Council, launched in 2025 to guide PAI’s work on AI’s societal impacts, is helping connect these conversations to broader policy efforts worldwide. In short, the code invites ongoing stakeholder input to inform how the EU AI Act will operate in practice, with an emphasis on transparency as a public good rather than a mere checkbox.

The guidance makes clear that no single transparency method is perfectly robust. Instead, policymakers and industry researchers argue for a mix of approaches that can illuminate how content was produced without compromising privacy. The underlying intent is straightforward: when audiences can see who produced a piece of AI-generated audio or video, and how it was edited or deployed, they can make more informed judgments about what they’re watching or sharing. The framing emphasizes that the transparency toolkit should be layered and adaptable, reflecting fast-changing technology while safeguarding individual rights.

The urgency is not theoretical. The document cites real-world risks from AI-generated media, including manipulation of election perceptions, the creation of fake dating profiles, and misrepresentations of ongoing conflicts, such as the war in Iran. Those examples underscore why audiences need clearer signals about origin and editing processes, and why platforms, publishers, and technology providers must participate in a credible transparency regime.

Industry observers offer practical takeaways. First, there is a tension between transparency and privacy. Disclosures must be meaningful without revealing sensitive data or enabling misuse, so privacy-by-design considerations have to be baked into every disclosure mechanism from the start. Second, the voluntary nature of the Code raises concerns for smaller players who may lack the resources to implement rigorous disclosures. The risk is a two-tier system where large platforms set the pace, leaving smaller providers behind unless safeguards and scaled tools are available. Third, expectations span borders. The EU is not alone in pursuing transparency; policymakers in other regions are moving similarly, which means cross-border compatibility and the potential for divergent standards need careful management. Fourth, there is a gap between “tick-box” compliance and real understanding. Clear definitions of what counts as transparent, and practical workflows for verification and auditing, are essential to avoid superficial disclosures that audiences ignore.

What happens next will matter for compliance teams and tech leaders. The EU’s voluntary Code of Practice is a proving ground that could shape future, enforceable rules under the EU AI Act. Compliance deadlines, when they come, will depend on the Act’s regulatory timeline, and enforcement will rely on regulators bridging the Code to formal rules. In the near term, expect continued stakeholder engagement, more detailed guidance on how disclosures should appear to users, and ongoing testing of how transparency signals influence trust and understanding across diverse audiences.