EU moves to require labels on realistic AI-generated content
Brussels will force labels on realistic AI images, audio and video from Aug. 2, 2026, and its icons and code of practice may become the template for global platforms.
People in the European Union will have to be informed from Aug. 2, 2026, when they are interacting with AI systems or seeing certain AI-generated or manipulated content. Under Article 50 of the AI Act, the target is not AI itself but content that looks real enough to be mistaken for a photograph, a witness recording, a voice memo or a news clip.
The Commission has published guidelines on transparency obligations for providers and deployers of certain AI systems, and it has published a final Code of Practice on marking and labelling AI-generated content. Brussels has also released EU icons that creators, publishers and other deployers of generative AI systems may use when they label synthetic material under the transparency rules.
The Commission’s guidance applies to providers and deployers, meaning the companies building AI systems and the organizations putting them into use will both carry obligations. The guidelines define the scope of Article 50, and the FAQ clarifies the concepts and duties that will govern disclosure. A consultation on draft guidelines was opened before the rule takes effect, giving platforms and developers a limited window to adjust product design, upload flows and moderation systems.
In practice, “authentic-looking” means media that can plausibly pass as human-made or camera-captured even when it has been created or materially modified by AI. That includes deepfake video, cloned voices, altered photographs, manipulated documents and synthetic advertising images, especially in elections, emergencies and conflict, where manipulated material can move fast and cause real-world harm.
The compliance burden will fall hardest on social platforms, news organizations, advertisers and AI developers that distribute content at scale. They will need ways to verify provenance, attach labels consistently and surface disclosures in a way users can actually notice. The Commission’s code of practice is voluntary soft law.
A 2026 legal note from Lewis Silkin found the labeling issue can trigger consumer backlash, citing the response to AI-generated product photographs used by J.Crew.
This article was produced by Prism’s automated news system from verified source data, official records, and press releases, then run through automated quality and moderation checks before publishing. The system is built and supervised by the people who set the standards it runs under. Read our full AI policy.
Did this article answer your question?
