EU publishes AI Act Article 50 transparency guidelines ahead of August deadline

The European Commission published guidelines on July 20 explaining how providers and deployers of AI systems must meet the transparency obligations in Article 50 of the EU AI Act. The rules take effect August 2.

The guidance arrives two weeks before the deadline and spells out duties that have been vague in practice. Providers must tell users plainly when they are interacting with an AI system, and embed machine-readable markers so that AI-generated or manipulated content can be detected downstream. Deployers face their own line: they must disclose deepfakes, flag AI-generated text on matters of public interest published without human editorial review, and reveal when emotion-recognition or biometric-categorization systems are in use.

“These guidelines help providers and deployers to comply with their obligations under the AI Act, while also helping citizens to know when they are interacting with AI,” said Executive Vice-President Henna Virkkunen.

The Commission paired the document with a Code of Practice on Transparency of AI-Generated Content, an EU-designed icon set for labeling AI output, and a support-service helpdesk. A separate labeling and tracking compliance deadline runs to December 2.

Guidelines are not binding law, and much of the enforcement detail – how regulators will test whether a watermark is genuinely machine-readable, or what counts as adequate disclosure – remains to be worked out. For companies shipping generative AI into the EU, though, the August 2 date is now the one that matters.