What happened
On 28 October 2026, at the Virtual Leaders' Meeting on Frontier AI Safety co-hosted by Norway and Finland, Singapore Prime Minister Lawrence Wong proposed that interested parties and countries launch a voluntary international pilot on AI incident reporting. He suggested agreeing a common set of incident categories and information to be shared, exploring secure collection/exchange, and drawing on lessons from international aviation incident reporting. He recommended organising cooperation around specific workstreams with smaller cross-regional groups and building on existing efforts such as Singapore's International Scientific Exchange on AI Safety, the Singapore Consensus on Global AI Safety Research Priorities, and the UN Independent Scientific Panel supporting the UN Global Dialogue on AI Governance.
Why it matters
This is a head-of-government proposal that could lead to a shared, cross-border AI incident-reporting mechanism — a precursor to a wider reporting framework. It raises expectations that frontier-model failures and near-misses will be shared systemically rather than left to developers, shifting transparency and recordkeeping norms for deployers and governments.
Action needed
Monitor Singapore's convening of the pilot and its proposed incident categories. Organisations deploying frontier models should anticipate maintaining incident/near-miss records and reporting channels that could be shared under a future pilot.