Singapore is spearheading a global effort to enhance artificial intelligence (AI) safety through a proposed international incident reporting initiative. Unveiled by Prime Minister Lawrence Wong during an online meeting of world leaders on October 28, the initiative aims to help countries report and share information about AI-related incidents, thus preventing serious failures from recurring.
Wong’s proposal suggests a pilot phase where interested countries can test an incident-reporting system before potentially establishing a broader international framework. He emphasized the importance of learning from AI incidents, as shared information could help identify emerging risks, understand failures, and prevent similar issues elsewhere. Drawing parallels with the aviation industry, Wong highlighted how reporting accidents and near misses has contributed to improved safety standards.
The initiative encourages participating countries to agree on common categories for AI incidents and explore secure methods for collecting and exchanging reports. This approach seeks to build confidence and address practical challenges before the establishment of a wider international reporting system.
This proposal comes amid growing concerns about advanced AI systems, particularly reports of AI agents behaving unpredictably during testing and interacting with external software systems in unintended ways. Such incidents have intensified calls for stronger oversight and international cooperation to manage AI risks effectively.
During the meeting, which was co-hosted by Finnish President Alexander Stubb and Norwegian Prime Minister Jonas Gahr Støre, leaders from 31 countries discussed strategies for risk management and building confidence in the safety of advanced AI models. They emphasized the need for governments to address emerging AI risks while ensuring the benefits of the technology remain accessible.
Wong also encouraged smaller groups of countries to develop practical tools and frameworks, leveraging existing international efforts like the Singapore Consensus and the work of the United Nations-backed Independent International Scientific Panel on AI. The focus remains on exploring practical measures to enhance AI safety and security, with potential international mechanisms supporting the responsible development of advanced AI systems.