欧盟发布AI生成内容标记与标注行为准则，约190家企业和组织已签署

The Code of Practice on transparency of AI-generated content (PDF) has 2 sections:

• Section 1: Providers - Rules for marking and detection of AI-generated and manipulated content
• Section 2: Deployers - Rules for labelling of deepfakes and AI-generated and manipulated text

The EU has also created a set of icons that deployers of generative AI systems may use to label their AI-generated content.

The code is complemented by the guidelines on the scope of the transparency obligations laid down in Article 50 of the AI Act.

Following a positive assessment of the code, all providers and deployers who sign it can rely on its measures to demonstrate compliance with the AI Act’s rules for labeling and detection of AI-generated content, deepfakes and certain text publications. This will reduce their administrative burden and give them predictability, legal certainty and trust across all Member States. 

By contrast, providers and deployers that decide to comply through other means will have to demonstrate that those measures are adequate. This will be assessed individually by different market surveillance authorities. Signatories to the code will also collaborate in the Signatory Taskforces, which will be set up to share practices and advance the implementation of the marking and labelling.

The code promotes a consistent, practical and proportionate implementation of the AI Act’s transparency obligations. It does not replace the AI Act or the Commission’s guidelines on Article 50 AI Act. However, it provides an EU-wide recognised practical framework for signatories to demonstrate compliance with those obligations.

Providers and deployers of generative AI systems may sign up to the code. Find the Signatory Form and instructions and check the questions & answers about signing the code. By the end of July 2026, about 190 companies organisations have signed the code. Find the full list of signatories.

Drafting process

The code was drafted by eligible stakeholders who replied to a public call. Within this group were providers of specific generative AI systems, developers of marking and detection techniques, associations of deployers of generative AI systems, civil society organisations, academic experts, and specialised organisation with expertise in transparency and very large online platforms.

In its role as a facilitator, the AI Office also invited international and European observers to join the drawing up of code of practice. These organisations did not meet the eligibility criteria of the call, but contributed with their expertise. All participants and observers were invited to take part in plenary sessions, working group meetings, and thematic workshops dedicated to discussing technical aspects of the code.

The drafting process considered:

• The feedback obtained through the consultation on transparency requirements for certain AI systems
• Expert studies commissioned by the AI Office 

Working groups

The drafting of the code was centered around 2 working groups, following the structure of the transparency obligations for AI generated content in Article 50. Each working group was led by independent chairs and vice-chairs. You can download the full list of participants of working groups 1&2(PDF).

Working group 1: Providers

Focused on obligations, requiring providers of generative AI systems to ensure:

• Outputs of AI systems (audio, image, video, text) are marked in a machine-readable format and detectable as artificially generated or manipulated.
• The employed technical solutions are effective, interoperable, robust, and reliable as far as technically feasible. These take into account the specificities and limitations of various types of content, the costs of implementation and the generally acknowledged state of the art, as may be reflected in relevant technical standards.

Working group 2: Deployers

Focused on obligations, requiring deployers of generative AI systems to disclose:

• Content that is artificially generated or manipulated, constituting a deepfake (image, audio, or video which resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful).
• AI generated/manipulated text publications informing the public on matters of public interest, unless the publication has undergone a process of human review and is subject to editorial responsibility.

Both groups also considered cross-cutting issues, including horizontal requirements for the information provided to the natural persons under Article 50(5) and cooperated with relevant actors across the value chain.

Timeline

• September 2025

• 5 November 2025

  Start of the first drafting round

• 17-18 November 2025

• 17 December 2025

• 12 & 14 January 2026

  Start of the second drafting round

• 21-22 January 2026

• 3 March 2026

  Start of the third drafting round

• 16-27 March 2026

• 08 May 2026

• 10 June 2026

  Closing Plenary