EU AI Act content-transparency duties take effect August 2

The European Commission says the transparency obligations in Article 50 of the EU AI Act apply from August 2, 2026. The rules cover several distinct situations rather than imposing one universal badge on every AI-assisted output.

Providers of AI systems that interact directly with people must disclose that users are dealing with a machine unless that fact is already obvious. Providers of systems generating synthetic audio, image, video or text must also support machine-readable marking of covered outputs. Deployers have additional disclosure duties for deepfakes and certain AI-generated or manipulated text published to inform the public.

The Commission says the public-interest text disclosure does not apply where the content has undergone human review or editorial control and a person or organization accepts editorial responsibility. The precise facts and workflow therefore matter; using any automated assistance is not, by itself, the complete legal test.

The Commission has published optional EU icons for labeling covered AI-generated content and a voluntary Code of Practice on Transparency of AI-Generated Content. The code separates provider practices for marking and detection from deployer practices for labeling deepfakes and covered text.

The Register independently reported the August deadline and the disclosure requirements for chatbots, agents and synthetic content. Organizations choosing methods outside the voluntary code still need to demonstrate that their controls satisfy the binding rules.

For practitioners, this is systems work, not merely copywriting. Output metadata, interface disclosures and editorial-review records may span model APIs, content-management systems, agent interfaces and media pipelines. A control that exists in one product surface but disappears after export or republishing will not provide reliable provenance.

Teams should map which outputs fall within each Article 50 category, where machine-readable marks are attached, how visible disclosures are presented, and how human editorial decisions are recorded. Legal interpretation should be reviewed with qualified counsel; the retrieved sources establish the Commission's implementation guidance and timing, not organization-specific legal advice.

Useful next signals include national market-surveillance guidance, adoption of the code and icons, and technical details on how provenance marks survive transformation across tools. Participation in the voluntary code and legal applicability remain separate questions.