Singapore has aligned with the European Union and 19 other countries in advocating for enhanced international oversight and safety measures for frontier artificial intelligence (AI) models. This initiative was formalized in a joint declaration during the United Nations General Assembly in New York.
The declaration emphasizes the necessity of international collaboration to harness AI’s benefits while mitigating risks associated with powerful AI systems. Singapore’s Prime Minister Lawrence Wong underscored the importance of AI remaining under human control and oversight, highlighting that the risks posed by AI transcend national borders.
The participating countries are urging UN member states to expand on existing international frameworks and consider establishing a dedicated international institution for frontier AI. This body would be tasked with setting standards, supporting verification processes, and coordinating responses as AI systems achieve significant advancements.
Signatories of the declaration, which include Singapore as the sole Southeast Asian participant, are calling for governments to develop common AI standards, enhance transparency, and establish shared methods for reporting serious AI safety incidents. They also aim to increase access to scientific expertise and trusted AI evaluation capabilities globally.
The declaration appeals to AI companies to bolster their safety measures by rigorously testing systems before deployment and employing independent evaluations to assess potential risks. Notably, the United States and China, prominent players in the AI field, have not endorsed the declaration thus far. The call for stronger AI safeguards remains open to additional countries willing to support this international effort.