Home / Guidance

Guidance registry

Everything official that interprets or completes the Act — Commission guidelines, codes of practice, delegated and implementing acts, consultations — with what's adopted, what's pending, and what's due when.

TYPEINSTRUMENTSTATUSDATE

AMENDMENT Digital Omnibus on AI Amends the AI Act — deadlines, registration, AI Office scope, new prohibition. Adopted EP 16 Jun and Council 29 Jun 2026; final act signed 8 Jul 2026. adopted; OJ number + CELEX pending publication OJ PENDING

GUIDELINE Guidelines on Article 50 transparency obligations Commission guidance on the Art. 50 transparency duties - chatbot disclosure, machine-readable marking of synthetic content, deepfake and emotion-recognition labelling. Draft published 8 June-window consultation (8 May to 3 June 2026); final guidelines published 20 July 2026, ahead of the Art. 50 duties applying on 2 August 2026. The Art. 50(2) marking duty has a transition to 2 December 2026 for systems already on the market. Complements the Code of Practice on Transparency of AI-Generated Content. adopted JUL 20, 2026

CODE Code of Practice on Transparency of AI-Generated Content Final text published 10 June 2026. Sets marking and detection practices for providers of generative systems, and deepfake plus public-interest-text labelling for deployers, supporting Art. 50. Signatory uptake and treatment of non-signatories are still developing. final JUN 10, 2026

GUIDELINE Guidelines on high-risk classification Draft Commission guidelines on applying Art. 6 and Annex III, with worked examples of systems that are and are not high-risk. Published 19 May 2026, missing the original Art. 6(5) deadline of 2 February 2026; feedback open until 23 July 2026. A final version is expected before the standalone high-risk duties apply on 2 December 2027 under the Digital Omnibus. Treat the current text as provisional. consultation DUE DEC 2, 2027

GUIDELINE Guidance and template — serious incident reporting (Art. 73) Commission draft guidance under Art. 73(7), with a reporting template, on how providers of high-risk AI systems report serious incidents to market surveillance authorities. Consultation closed 7 November 2025; the final version was expected around 2 August 2026, and the underlying Art. 73 duty now applies from 2 December 2027 under the Digital Omnibus. This is guidance, not an implementing act. A separate published template already covers serious incidents involving GPAI models with systemic risk. The 2025-09-26 draft date is approximate and should be re-checked against the consultation page. consultation DUE DEC 2, 2027

GUIDELINE AI Office template — training-data summary Mandatory template for the public summary of training content that general-purpose AI model providers must publish under Art. 53(1)(d), published with an explanatory notice on 24 July 2025. The duty applied from 2 August 2025; models already on the market before that date have until 2 August 2027. The AI Office can verify compliance and issue corrective measures from 2 August 2026. adopted JUL 24, 2025

GUIDELINE Guidelines for GPAI providers AI Office guidance on the scope of the general-purpose AI model rules (Art. 51 to 55): the compute threshold for systemic risk, who counts as a provider, and available exemptions. Complements the GPAI Code of Practice. adopted JUL 18, 2025

CODE GPAI Code of Practice Voluntary route for general-purpose AI model providers to show they meet Art. 53 and 55 duties on transparency, copyright, and safety and security. Final text published by the AI Office on 10 July 2025; the Commission and the AI Board confirmed its adequacy on 1 August 2025 and the initial signatory list was published the same day, ahead of the GPAI obligations applying from 2 August 2025. The AI Office maintains and keeps expanding the public signatory list. adopted JUL 10, 2025

GUIDELINE Guidelines on the definition of an AI system Commission guidance clarifying what counts as an AI system under Art. 3(1), to help operators decide whether the AI Act applies to a given product at all. Interpretive, not binding. adopted FEB 6, 2025

GUIDELINE Guidelines on prohibited practices Commission guidance on the scope of each practice banned under Art. 5, adopted alongside the February 2025 start of the prohibitions. Interpretive only; the binding text remains Art. 5 of the AI Act. adopted FEB 4, 2025

DELEGATED Annex I sectoral delegated acts Under the Digital Omnibus on AI (adopted by Parliament 16 June 2026 and Council 29 June 2026), high-risk duties for AI embedded in Annex I regulated products are deferred to 2 August 2028, the definition of 'safety component' is narrowed, and the Commission is empowered to adopt delegated acts amending Annex III of the Machinery Regulation with AI-specific health and safety requirements by 2 August 2028. No such delegated act has been adopted yet. pending DUE AUG 2, 2028

IMPLEMENTING Implementing act — post-market monitoring plan template Art. 72(3) requires the Commission to adopt an implementing act setting a template for the post-market monitoring plan that high-risk providers must maintain, by 2 February 2026. That deadline passed with no adopted act as of July 2026. The Commission's 2026 guidelines roadmap instead foresees guidelines offering a voluntary post-market monitoring template during 2026. High-risk duties themselves now apply from 2 December 2027. pending DUE FEB 2, 2026

New instruments detected via the daily OJ feed and AI Office news Sources: digital-strategy.ec.europa.eu · EUR-Lex · verified 2026-07-16