SINGAPORE: Singapore is studying whether new safety measures are needed in high-risk uses of artificial intelligence, as increasingly capable and autonomous AI systems create new risks that existing safeguards may not fully address.

In essential services and other higher-risk AI uses, these safeguards could include more rigorous testing, independently verifiable evidence that safety measures work, tighter controls on deployment and stronger regulatory oversight, Minister for Digital Development and Information Josephine Teo said in a written parliamentary response on Tuesday (Oct 6).

She added that Singapore would continue to enable AI innovation while taking a “risk-based and multi-layered approach”, which involves setting limits on AI deployment according to risk.

For example, before deploying an AI agent in the public service, the government considers what systems and data it can access, what it is allowed to do, and the consequences of unintended or unauthorised actions, said Mrs Teo.

Singapore’s approach is to deliberately limit what AI agents can do and to presume that malicious actors will use AI to cause harm more cheaply, quickly and at scale, she said.

She added that there has been no reported cyber attack against government agencies involving an unsupervised AI agent so far.

The Cyber Security Agency of Singapore (CSA) and Government Technology Agency of Singapore (GovTech) use AI to strengthen the security of government systems and critical information infrastructure, said Mrs Teo. This includes identifying vulnerabilities and potential threats more quickly.

CSA has also strengthened cybersecurity requirements for operators of critical information infrastructure, shared intelligence on active threats for them to take pre-emptive action, and given practical guidance to strengthen cyber defences and secure AI systems, she said.

The government is also looking into how existing channels for reporting cyber incidents through CSA, GovTech and sector leads can be better leveraged to identify incidents involving AI, support remediation and improve safeguards.

It is also looking into whether further coordination or reporting arrangements are needed, she said.

Mrs Teo said the government does not dismiss “catastrophic or even extinction-level” AI risks, nor assume that they will materialise.

Its approach is to monitor evidence closely, build technical capabilities to understand advanced AI systems, and work internationally on measures to address severe risks as the evidence develops.

To that end, the National AI Council provides the strategic direction on the nation’s broader approach to AI risks. Singapore also recently backed an international call by more than 20 world leaders for stronger safeguards around frontier AI.

Mrs Teo said the government also engages frontier AI developers for information and access to their models where appropriate.

“We currently do not make this a requirement, because access to a model by itself does not necessarily give a complete picture of the risks,” she said.

“A rigid requirement could even be counter-productive if it leads companies to limit their cooperation or information-sharing.

“Our preference is therefore to build strong collaborative relationships with frontier AI labs that encourage substantive and timely information-sharing.”Mrs Teo added that as AI becomes more capable of acting on behalf of people, more robust safeguards must correspondingly be introduced, and the ability to understand emerging risks and respond to problems must be strengthened.

“This will help build trust in AI as a technology that serves the public good,” she said.