Singapore proposes UN framework convention on AI safeguards

Foreign Minister Vivian Balakrishnan proposes shared AI testing, cross-border incident reporting and a possible UN convention. No treaty has been adopted.

Published

Sep 27, 2026

Signal type

Policy warning

Confidence

100%

Organization

Singapore Ministry of Foreign Affairs, United Nations

Sources

Last reviewed

Sep 29, 2026

Singapore's Foreign Minister Vivian Balakrishnan has proposed considering a UN framework convention on AI safeguards. He spoke at the General Assembly in New York on September 26, 2026; the foreign ministry published the speech the next day.

Balakrishnan explicitly identifies loss of control over autonomous systems as a risk. He calls for rigorous testing before deployment, clear limits on permitted actions and ways for humans to intervene. Countries should develop comparable tests and be able to report serious AI incidents quickly across borders.

The minister also suggests considering an international institution for technical standards and verification. At the same time, he regards a general halt to development as too late given geopolitical competition and financial incentives. His proposal pairs continued development with shared safeguards.

No convention has been adopted through this speech. In a follow-up media interview, Balakrishnan says it is too early to conclude whether one is necessary. This Risk Signal records a specific diplomatic initiative, not new operational oversight or an individual AI incident.