Every AI Policy Ever · open in the interactive register

Published (non-binding, current)Non-binding · Report (government-commissioned)

Preliminary Report of the Independent International Scientific Panel on AI: Evidence-based assessment of opportunities, risks and impacts of AI

Record UN-08

Issued by
Independent International Scientific Panel on AI (co-chairs Yoshua Bengio and Maria Ressa)
Jurisdiction
United Nations
Level · region
International · Global
Published / adopted
2026-07-01
Language
Arabic, Chinese, English, French, Russian, Spanish

Summary

The first report of the Independent International Scientific Panel on AI, which the General Assembly established by resolution 79/325, assessing the capabilities, opportunities, risks and impacts of AI. Its central warning is that current safeguards cannot keep pace with the growth of AI capabilities, and it describes an evidence dilemma: by the time the evidence exists, it may be too late to act. It reports findings in seven domains, including security, human rights and democracy, and child safety, and finds that reliable methods for retaining control over highly autonomous AI systems are lacking and that evaluation methods remain underdeveloped.

Safety topics

general-purpose / frontier AIcatastrophic riskevaluation & testingchild safetyresearch prioritiesinternational cooperation

Status notes

Released on 1 Jul 2026 to inform the inaugural Global Dialogue on AI Governance in Geneva (6-7 Jul 2026). Mandated by A/RES/79/325. The Panel calls it the beginning of its work, to be updated progressively with thematic briefs on emerging issues, and says its next annual report will inform the second Global Dialogue in New York in May 2027; the Panel's site lists a first thematic brief, on AI agents and the risk of losing human control (UN-09). The report states that it represents a broad consensus of the Panel's members, who serve in their personal capacities, and does not represent the views of the United Nations.

Sources

Open the official source https://www.un.org/independent-international-scientific-panel-ai/en/preliminary-report

Secondary source: https://www.un.org/independent-international-scientific-panel-ai/sites/default/files/2026-07/en_executive_summary.pdf

Local copy

Archived. Last attempt 2026-10-17 13:37:53 UTC

Official source (third-party copy, not published) archive/UN-08/2026-10-11/official.html
Retrieved 2026-10-11 13:57:51 UTC
SHA-256 139e78feee2420fa65057c252dbdca4c6da44004078b56fec76054686a05fedc
Linked document (PDF) (third-party copy, not published) archive/UN-08/2026-10-11/document-1.pdf
Retrieved 2026-10-11 13:57:59 UTC
SHA-256 7939c552f36bdc05896b90f512da3faefe32987a2d5cb83f6ee0a0f2c673ba31
Linked document (PDF) (third-party copy, not published) archive/UN-08/2026-10-11/document-2.pdf
Retrieved 2026-10-11 13:58:06 UTC
SHA-256 68bde9e9c5573ec2f404d1221cd87db1a4fdd6c5f9619f7950dffc49fcc79fe7
Secondary source (third-party copy, not published) archive/UN-08/2026-10-11/secondary.pdf
Retrieved 2026-10-11 13:58:11 UTC
SHA-256 e63a0a429524761f4116a7be121c9b6cff80a3d5cdbdf5a6f6c979a5defd021f

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 copies (un.org, 11 Oct 2026; the archiver found them unchanged on 17 Oct 2026): the Panel's Preliminary Report page, the full report PDF (59 pages), the Co-Chairs' message dated 1 July 2026 and the executive summary PDF, stored as the record's secondary source. The page confirms the title, the release materials dated 01 July 2026, the six language versions of the report and the seven domains; the full report confirms the title, the July 2026 date, the Panel's authorship and its establishment by resolution 79/325, the disclaimer and the findings in the summary. The stored page lists no updated version of the report. The summary no longer speaks of an 'evidence timing problem', a phrase the report does not use; it calls this an evidence dilemma.
Second check: Independent check by Claude Code (claude-sonnet-5-5), 2026-10-17: 6 field(s) supported.

History

  1. 2026-10-17notes: Launched 1 Jul 2026, shortly before the first Global Dialogue session in Geneva → Released on 1 Jul 2026 to inform the inaugural Global Dialogue on AI Governance — Added what the report says about its own updating, the next annual report, the thematic brief and its disclaimer; all from the stored page and report.2026-10-17-recheck-international-1 · Claude Code (claude-opus-5-5)
  2. 2026-10-17safety_topics: ["general-purpose / frontier AI", "child safety", "research priorities", "intern → ["general-purpose / frontier AI", "catastrophic risk", "evaluation & testing", " — Added 'catastrophic risk' (the gap between capabilities and risk management 'may lead to catastrophic outcomes') and 'evaluation & testing' (evaluation methods underdeveloped; independent assessment institutions embryonic).2026-10-17-recheck-international-1 · Claude Code (claude-opus-5-5)
  3. 2026-10-17summary: The UN Scientific Panel's first assessment of AI opportunities, risks and impact → The first report of the Independent International Scientific Panel on AI, which — The phrase 'evidence timing problem' is not in the report (it speaks of an 'evidence dilemma'); rewritten from the stored report, with the control and evaluation findings of its executive summary.2026-10-17-recheck-international-1 · Claude Code (claude-opus-5-5)
  4. 2026-10-17language: English, French, Spanish → Arabic, Chinese, English, French, Russian, Spanish — The page offers the full report, executive summary and message in all six UN languages (EN, AR, ES, FR, RU, ZH), not only English, French and Spanish.2026-10-17-recheck-international-1 · Claude Code (claude-opus-5-5)
  5. 2026-10-14language: English, French, Spanish (other UN languages) → English, French, Spanish — The field held a remark after the language name; it now gives the language names only (AGENTS.md 5.6).manual-2026-10-14 · Claude Code (claude-opus-5-5)

Related records

UN AI Scientific Panel & Global Dialogue resolution United Nations · Published (non-binding, current)

UN AI Panel brief on AI agents & loss of control United Nations · Published (non-binding, current)

International AI Safety Report 2026 International AI Safety Report (AI Safety Summit process) · Published (non-binding, current)

UN Global Dialogue on AI Governance 2026 Co-Chairs' Summary United Nations · Published (non-binding, current)

Citation

Independent International Scientific Panel on AI (co-chairs Yoshua Bengio and Maria Ressa). Preliminary Report of the Independent International Scientific Panel on AI: Evidence-based assessment of opportunities, risks and impacts of AI. 2026-07-01. https://www.un.org/independent-international-scientific-panel-ai/en/preliminary-report. Status: Published (non-binding, current); verified 2026-10-17 (level A). Local copy archive/UN-08/2026-10-11/official.html retrieved 2026-10-11 13:57:51 UTC, SHA-256 139e78feee2420fa65057c252dbdca4c6da44004078b56fec76054686a05fedc. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record UN-08.

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.