The Pulse
Singapore Urges UN to Explore an AI Safeguards Convention
Singapore Foreign Minister Vivian Balakrishnan told the UN General Assembly that countries should explore a UN Framework Convention on AI Safeguards, alongside rigorous testing before deployment, clear limits on autonomous systems and share
AI.info Team · Written with AI · fact-checked ·
Singapore Foreign Minister Vivian Balakrishnan used his Saturday address to the United Nations General Assembly to call for countries to explore a framework convention on artificial intelligence safeguards. The proposal would put shared international rules alongside national and technical measures, rather than leave each government to address cross-border risks alone.
Balakrishnan puts the UN at the center
Balakrishnan framed AI rules as the third of three priorities for keeping multilateralism fit for purpose, after consistent compliance with international law and reform of existing institutions. The UN, he argued, is the only apex international institution with universal membership. His address said any durable approach should give all states a meaningful stake, not only those with the most advanced models, the deepest pockets or the largest set of armaments.
“We must explore all options, including the idea of an UN Framework Convention on AI Safeguards.”
Vivian Balakrishnan, Minister for Foreign Affairs, Singapore
The wording presents a convention as an option to examine, not an agreed treaty or a settled negotiating plan. Speaking to Singapore media after the speech, he said: “There have been a variety of ideas floated, and we also offered the possibility of considering whether it is necessary to have a UN Framework Convention on AI Safeguards. Too early to conclude yet, but at least the discussions have begun.”
Testing, limits and human control
The address set out three broad areas of concern: loss of control over autonomous systems; the use of AI by rogue actors to create weapons of mass destruction or mass disruption; and emerging economic, social and political disruptions. Balakrishnan ruled out a pause: “It is too late to call for a halt, especially given the superpower rivalry and the huge financial incentives for companies that already have frontier models.” Instead he called for rigorous testing and evaluation before deployment, clear limits on autonomous systems, and mechanisms to intervene when systems behave in ways that were not intended. He said people must remain ultimately in control and accountable.
Balakrishnan compared the need for shared expectations to traffic lights, which drivers read the same way in New York, Beijing, Moscow or Singapore: countries need a common understanding of how to manage risks that do not stop at national borders. Trust is the scarce resource, he said, because states must believe that risks will be disclosed, testing will be credible and cooperation will not serve one country’s advantage.
He proposed starting with the science. Even states that differ on governance, he said, could work together to understand the capabilities and risks of AI models, develop comparable methods for testing and evaluation, and report serious incidents quickly across borders. He pointed to the UN’s Independent International Scientific Panel on AI, the Global Dialogue on AI Governance and the ITU’s AI for Good Global Commission as bodies already doing part of that work.
A possible role for standards and verification
Singapore also raised the possibility of an international body carrying out some functions for AI that exist in other fields. Balakrishnan pointed to the International Telecommunication Union, which sets technical standards, and the International Atomic Energy Agency, which carries out verification. He framed it as an institution the world “may also need”, alongside the convention idea rather than as a finished design.
Singapore said it had joined a call for greater safety and security of frontier AI models led by Finland and Norway, and would keep contributing through the digital pillar of the Forum of Small States and by making its governance frameworks and testing tools available to all. The proposal remains at the discussion stage; the minister said that “the work of building trust and laying the ground for global cooperation must begin now.”