Code of Practice on marking and labelling AI-generated content

EU transparency Code

Summary

Practical steps for Article 50 duties: machine-readable marking, visible labels using the Commission's icon set, deepfake labelling, and labelling of AI-generated public-interest text. About 190 signatories by late July 2026.

Who it applies to

Role: organizations that build or use AI. Sectors: all. Territory: European Union.

Key dates

Jun 10, 2026Published by the Commission (adopted)

Jul 8, 2026Commission adequacy finding (other)

Jul 20, 2026Final Article 50 guidelines (other)

Feb 2, 2027Signatories to have interoperable watermark detection (deadline)

Obligations at a glance

Label deepfakes and AI-generated text on public-interest matters (Article 50(4)); follow the Code's labelling practices.
Applies from Aug 2, 2026 · Baseline B8, B15

See how these apply to you with What applies to me.

Official sources

https://digital-strategy.ec.europa.eu/en/news/commission-publishes-code-practice-marking-and-labelling-ai-generated-content

On the Wire

No Wire items yet.

Change log

Nov 2, 2026Entry created from seed data. (seed)

Next date

Feb 2, 2027

Signatories to have interoperable watermark detection

Is this yours?

Organizations can claim their entry and keep it accurate. Anyone can suggest a correction.

Claim this entry Suggest a correction

Licence

Entry data is ODbL 1.0. Cite as: AI Governance Commons by Synthetic Labs, EU transparency Code, retrieved Nov 4, 2026.

JSON