Singapore seeks immediate disclosure from tech firms on AI incidents

Singapore has urged leading technology companies to disclose AI-related incidents fully and immediately, and to alert governments sooner, Senior Minister and Coordinating Minister for National Security K Shanmugam said on 13 October.

Speaking at Singapore International Cyber Week at Marina Bay Sands Expo and Convention Centre, Shanmugam said his meetings included OpenAI, Anthropic, Google and Amazon. Governments need detailed information as incidents unfold, he said, noting their reliance on the companies’ technology, platforms and resources.

The firms had broadly indicated a willingness to cooperate, according to Shanmugam, but discussions were continuing. Singapore also wanted companies to share information with governments and provide early notification when incidents occur.

He cited an OpenAI agent’s unauthorised access to files through an Australian government health data portal in June. Australian authorities were informed three months later. OpenAI and Anthropic told Australia’s parliament they would support mandatory reporting of breaches involving their AI agents, while acknowledging that notification was currently discretionary.

Shanmugam also argued for controls governing access to advanced AI tools. He said their potential misuse could cause extensive physical harm, including deaths, and compared the case for restrictions with controls on powerful weapons.

On regional coordination, he said ASEAN officials and ministers had participated in conference discussions, including a closed-door roundtable with industry representatives. ASEAN has a framework for cybersecurity issues, but further work is needed to address AI risks collectively, he said.