The Code of Practice on AI-generated content.
Reviewed
Short answer
The Code of Practice on Transparency of AI-generated Content is a voluntary code, published on 10 June 2026, that the Commission and the AI Board assessed as adequate for showing compliance with the marking and labelling duties in Article 50(2), (4) and (5) of the AI Act. Section 1 is for providers of generative AI (machine-readable marking and detection), Section 2 for deployers (labelling deepfakes and certain text), and about 190 organisations had signed by the end of July 2026.
Official sources: European Commission: Code of Practice on Transparency of AI-generated Content · Commission news, 31 July 2026
The two sections
| Section | For | Main commitments |
|---|---|---|
| 1. Marking and detection | Providers of generative AI systems; may also be signed by model providers and providers of marking and detection technology | At least two layers of machine-readable marking: digitally signed, time-stamped metadata and an imperceptible watermark (for free-form text a watermark alone, applied to text longer than 200 tokens); do not remove marks; a detection tool free of charge; effectiveness, reliability, robustness and interoperability; testing, training and cooperation with authorities |
| 2. Labelling | Deployers who publish deepfakes or AI-generated text on matters of public interest | An icon (the EU icon) or equivalent label with design and placement rules, for example at the top right of an image or video, at the start and after breaks in video, an audible disclaimer for audio; internal processes and staff awareness; a lighter regime for artistic works; conditions for human review of text |
On detection, the code says signatories make their detection solution available free of charge; signatories with fewer than 1,000,000 monthly users whose detection has substantial operating costs may charge a reasonable fee above a high per-user volume, but market surveillance authorities always get free, unlimited access.
Who has signed
The Commission reported about 190 signatories by the end of July 2026, from IT, telecoms, education and retail among others, about half of them small and recent companies. It named, for Section 1, Aleph Alpha, Anthropic, Black Forest Labs, Cohere, Google, Meta, Microsoft, Mistral, OpenAI and Synthesia, and for Section 2, Bulgari, Fastweb, Getty Images, Iberdrola, Lenovo and Lufthansa. Its list, last updated on 24 September 2026, showed 95 Section 1 signatories and 192 Section 2 signatories, and stays open for new ones.
How to sign
- Check eligibility: providers of generative AI systems (Section 1) and deployers subject to Article 50(4) (Section 2); model providers and marking or detection technology providers may sign Section 1.
- Download the signatory form from the Commission’s page and have it signed by someone with authority to bind the organisation, such as a senior executive.
- Send it to the AI Office mailbox given on the Commission’s page. Signing remains possible after the first deadline of 27 July 2026.
What signing does, and does not do
- For signatories, enforcement focuses on monitoring adherence to the code, which the Commission describes as more predictable and less burdensome.
- Adherence is not conclusive evidence of compliance (Commission opinion of 9 July 2026).
- Not signing is not a breach: non-signatories must show compliance by other adequate means, assessed by market surveillance authorities, and may face more requests for information.
- Under Article 50(7), as amended in 2026, the Commission assesses whether adherence is adequate and may adopt an implementing act with common rules if it finds a code inadequate.
- The AI Office plans to consider formal updates at least every two years.
Questions
- Is the Code of Practice mandatory?
- No. Signing is voluntary, while the Article 50 duties it covers are legal obligations. Providers and deployers who do not sign must demonstrate compliance through other adequate means.
- Can a small company sign?
- Yes. Any provider or deployer of a generative AI system subject to Article 50(2) or (4) may sign, and the Commission reported that about half of the first signatories were small and recent companies.
Sources
- Code of Practice on Transparency of AI-generated ContentEuropean Commission, Digital Strategyread
- Code of Practice on Transparency of AI-Generated Content (final text, PDF)European Commissionread
- Commission opinion on the assessment of the Code of Practice on Transparency of AI-generated Content (9 July 2026)European Commissionread
- Strong backing for the Code of Practice on Transparency of AI-generated Content (31 July 2026, with the list of signatories)European Commissionread
- How to sign the Code of Practice on transparency of AI-generated contentEuropean Commissionread
- Signing the Code of Practice on Transparency of AI-generated Content — questions and answersEuropean Commissionread
- Transparency obligations under Article 50 of the AI Act — questions and answersEuropean Commissionread
- AI Act, Article 50: Transparency obligationsEuropean Commission, AI Act Service Desk (AI Act Explorer)Consolidated text as at 27 July 2026 (includes the Digital Omnibus, Regulation (EU) 2026/1744), shown in the Commission’s AI Act Explorer.read
- Regulation (EU) 2026/1744 (Digital Omnibus on AI) — Article 111(4) transition, Article 50(7)EUR-Lex / Official Journal of the EURead on 2 November 2026 from the EU Publications Office’s copy of the Official Journal text (publications.europa.eu, same act); EUR-Lex blocks scripted fetches.read
Last reviewed by Volkan Çatak, Founder, Proofcan.
Not legal advice. This page explains official texts in plain words and links each one; check the sources, or a lawyer, before relying on it for a decision.