Every AI Policy Ever · open in the interactive register

Published (non-binding, current)Non-binding · Guidelines / framework

Ethics Guidelines for Trustworthy AI

Record EU-20

Issued by
European Commission
Jurisdiction
European Union
Level · region
Supranational · Europe
Published / adopted
2019-04-08
Language
English

Summary

Guidelines by the High-Level Expert Group on AI set up by the European Commission on what makes AI trustworthy: lawful, respecting all applicable laws; ethical, respecting ethical principles and values; and robust, technically and with regard to its social environment. They set seven key requirements: human agency and oversight; technical robustness and safety, including a fall back plan; privacy and data governance; transparency, including that people know they are interacting with an AI system; diversity, non-discrimination and fairness; societal and environmental well-being; and accountability, with auditability and redress. An assessment list helps verify that they are met.

Safety topics

human oversighttransparency & labellingaccountability & liabilityevaluation & testing

Scope

Borderline scope Report of an independent expert group convened and published by the Commission; existing scope_note kept.

Status notes

Written by the High-Level Expert Group on Artificial Intelligence (AI HLEG), an independent expert group that the European Commission set up in June 2018, and presented on 8 April 2019, after a first draft published in December 2018 drew more than 500 comments in an open consultation. The assessment list in the guidelines was piloted by the Commission from 26 June to 1 December 2019; on that basis the AI HLEG presented the final Assessment List for Trustworthy AI (ALTAI), a self-assessment checklist and the group's third deliverable, in July 2020 (made public on 17 July 2020). The guidelines are offered in 23 EU languages, together with a separate AI HLEG document on the definition of AI used for their purpose.

Sources

Open the official source https://digital-strategy.ec.europa.eu/en/library/ethics-guidelines-trustworthy-ai

Local copy

Archived. A source changed since the previous snapshot; both versions are kept. Last attempt 2026-10-17 18:20:51 UTC

Official source (third-party copy, not published) archive/_runs/2026-10-16-discover-leads/C002/2026-10-16/official.html
Retrieved 2026-10-16 11:40:35 UTC
SHA-256 f54a63e0a739de2f551a2974762e492852b85453c5c28b0282cc79928e19756e
Linked document (PDF) (third-party copy, not published) archive/EU-20/2026-10-17/document-1.pdf
Retrieved 2026-10-17 16:47:40 UTC
SHA-256 1780bb93aa16b0f0dc06490dceca9b217018b4742ae42e75c569131bd697305a
Linked document (PDF) (third-party copy, not published) archive/_runs/2026-10-16-discover-leads/C002/2026-10-16/document-2.pdf
Retrieved 2026-10-16 11:40:41 UTC
SHA-256 05d2831f581f4630e02b98b14464c8bb131434363c606d10ad4129bbf2f4b7df

Verification

A
The instrument's own official text, or an official register, was stored and confirms the title, date and status. checked 2026-10-17 · official-url-fetched
Stored and read: the Commission's library page for the guidelines on digital-strategy.ec.europa.eu (HTML and printable PDF, last updated 31 January 2024), which gives the title, presentation by the AI HLEG on 8 April 2019, the first draft of December 2018, the three components and seven key requirements, the piloting and ALTAI, and downloads in 23 languages; and the AI HLEG's ALTAI (July 2020), which says the group was set up by the Commission in June 2018, dates the guidelines to 8 April 2019 and the piloting to 26 June to 1 December 2019. The guidelines' own English text is not among this record's stored files, so claims that rest on it alone were removed. No stored source says the guidelines were withdrawn or replaced. The document carries no document number, so official_identifier is empty.
Second check: Independent check by Claude Code (claude-sonnet-5-5), 2026-10-17: 7 field(s) supported, fixes applied to scope_note.

History

  1. 2026-10-17scope_note: Written by the independent High-Level Expert Group on Artificial Intelligence th → Report of an independent expert group convened and published by the Commission; — pipeline run 2026-10-17-recheck-europe-1 (recheck), candidate EU-202026-10-17-recheck-europe-1 · Claude Code (claude-opus-5-5)
  2. 2026-10-17notes: Written by the High-Level Expert Group on Artificial Intelligence (AI HLEG), an → Written by the High-Level Expert Group on Artificial Intelligence (AI HLEG), an — rewritten from the stored files: '52 members', 'the first of its two mandated deliverables', the Commission disclaimer, '18 Dec 2018' and the fundamental rights impact assessment 'before development' are not in them (ALTAI gives the disclaimer and the impact assessment for ALTAI itself, and calls ALTAI the group's third deliverable); the piloting dates and ALTAI's publication come from ALTAI2026-10-17-recheck-europe-1 · Claude Code (claude-opus-5-5)
  3. 2026-10-17summary: Non-binding guidelines that define trustworthy AI as lawful, ethical and robust. → Guidelines by the High-Level Expert Group on AI set up by the European Commissio — rewritten from the stored library page: the four ethical principles and the list of critical concerns rest on the guidelines' own text, which is not among this record's stored files, and were removed (open_questions); the three components and seven requirements are given as the page gives them2026-10-17-recheck-europe-1 · Claude Code (claude-opus-5-5)
  4. 2026-10-16related: [] → [{"id": "EU-21", "slug": "eu-communication-on-ai-for-europe-2018"}] — two-way link to EU-21 added in run 2026-10-16-discover-leads-20182026-10-16-discover-leads-2018 · Claude Code (claude-opus-5-5)
  5. 2026-10-16Added to the register — found in pipeline run 2026-10-16-discover-leads, candidate C002 (EU/executive)2026-10-16-discover-leads · Claude Code (claude-opus-5-5)

Related records

EU Communication on AI for Europe (2018) European Union · Published (non-binding, current)

Citation

European Commission. Ethics Guidelines for Trustworthy AI. 2019-04-08. https://digital-strategy.ec.europa.eu/en/library/ethics-guidelines-trustworthy-ai. Status: Published (non-binding, current); verified 2026-10-17 (level A). Local copy archive/_runs/2026-10-16-discover-leads/C002/2026-10-16/official.html retrieved 2026-10-16 11:40:35 UTC, SHA-256 f54a63e0a739de2f551a2974762e492852b85453c5c28b0282cc79928e19756e. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record EU-20.

Improve this record

Know of an amendment, a status change or an error? Report news or an error on GitHub; the form opens with this record filled in.