AI transparency code: obligations and deadlines

Code of Practice on AI-generated content transparency and Article 50 guidelines

Regulator guidanceAI transparency code

Summary

The Code is voluntary and helps meet Article 50: it has a section for providers (technical marking) and one for deployers (labeling deepfakes and text). The guidelines explain which systems must disclose that they are AI, including virtual assistants, phone systems and agents. They do not mandate a single marking technique.

Who it applies to

Providers and deployers of chatbots, agents and generative systems.

Penalties

The rule has no penalty regime of its own, or it does not apply.

Official name: Code of Practice on Transparency of AI-Generated Content; Commission Guidelines on the transparency obligations under Article 50 of Regulation (EU) 2024/1689
Official reference: Código final publicado el 10 de junio de 2026; directrices de la Comisión sobre el art. 50 publicadas el 20 de julio de 2026 (C(2026) 5054)
Regulator: European Commission and European AI Office
Official source: Check the sourceHow we verify sources

Key dates

June 10, 2026Final Code published
July 8, 2026Commission finds the Code adequate
July 20, 2026Commission guidelines on Article 50

Obligations and dates

Obligations table
Obligation	Article or citation	Status	Date	Who it applies to	Verification
Align notices and labels with the Code and guidelines	Código de Buenas Prácticas (secciones 1 y 2); directrices del art. 50	Regulator guidance	Already applicable	If you use or sell chatbots, agents or AI-generated content.	Verified in the official source on November 4, 2026

Align notices and labels with the Code and guidelines

Código de Buenas Prácticas (secciones 1 y 2); directrices del art. 50Regulator guidance

Already applicable

When it relates: If you use or sell chatbots, agents or AI-generated content.
What the rule provides: Use visible, consistent labels (the Commission's icons can be used) for deepfakes and synthetic content.
Providers: document which marking technique is used and how it is detected.
Check that voice assistants and agents disclose at the start of the interaction.
Typical evidence: Internal labeling guide with examples.
Record of Code sign-up, if signed.
Related base documents: AI Interaction Notice and Synthetic Content Labeling
Supporting documents: External AI Communications Guide
Generative AI Intellectual Property Policy
Verification: Verified in the official source on November 4, 2026Checked on digital-strategy.ec.europa.eu: the final Code was published on Jun 10, 2026, with Section 1 (providers: marking and detection) and Section 2 (deployers: labeling), plus EU icons. The Article 50 guidelines were published on Jul 20, 2026 (C(2026) 5054). The page confirms the Commission and the AI Board found the Code adequate, but gives no date (Jul 8, 2026).