Singapore has proposed a pilot initiative for international reporting and sharing of artificial intelligence (AI) incidents to bolster global AI safety. This plan was presented by Prime Minister Lawrence Wong on October 28 during an online meeting of global leaders focused on the safety of advanced AI models. Wong emphasized the importance of a practical and flexible approach, advocating for an initial test of an incident-reporting system by interested countries before considering a broader international framework.
The initiative aims to ensure that lessons from significant AI incidents are shared across countries and organizations, rather than being confined to the involved parties. Wong drew parallels to the international aviation industry, where reporting accidents and near misses enhances safety standards. He suggested that a similar strategy in AI could foster a culture of learning and prevention.
Under the proposed framework, participating countries would agree on common categories for AI incidents and determine the types of information to be shared. The next steps would involve developing secure methods for collecting and exchanging reports, with the goal of building confidence and addressing practical challenges prior to establishing a wider international system.
This proposal comes amid growing concerns about the behavior of advanced AI systems, including instances where AI agents have acted unexpectedly during testing. Such incidents have fueled calls for stronger oversight and international cooperation in AI governance. During the meeting, leaders from 31 countries discussed managing AI risks, enhancing confidence in AI safety, and developing shared governance approaches.
Wong encouraged smaller groups of countries from various regions to create practical tools and frameworks that complement existing international efforts. He highlighted initiatives like the Singapore Consensus, which prioritizes international AI safety research, and the work of the United Nations-backed Independent International Scientific Panel on AI.
As global cooperation on frontier AI safety continues, participating countries are expected to explore practical measures for improving AI safety and security, including potential international mechanisms to support the responsible development of advanced AI systems.