S'pore calls for greater safety & security of frontier AI models, urges for global cooperation at UN General Assembly

In his delivery of Singapore's national statement at the UN General Assembly on Sep. 26, Minister for Foreign Affairs Vivian Balakrishnan urged the world to work together to develop safeguards for Artificial Intelligence (AI) models so that humans remain in control and accountable for AI.

"Let me give you an extreme thought experiment. Do you want the nuclear button to be in the hands of a human being or an AI? Think about that."

AI, which was mentioned several times at the UN General Assembly, is "arguably the most urgent frontier" for countries to formulate new rules and create new institutions.

The need for common rules to manage AI risks

"Its capability is advancing faster than our appreciation of its risks," Vivian said before giving examples of such risks.

These include the loss of control over autonomous systems and the possibility of rogue actors using AI to create mass destruction or disruptions to society.

Vivian said rigorous testing and evaluation of AI models are necessary before deployment. The world also needs clear limits on what autonomous systems can do and interventions when AI behaves in unintended ways.

Using the analogy of driving a high-performance car, the use of AI also requires good "brakes" and effective shared rules to ensure safety.

"Since the risks from AI do not remain confined behind borders, it means we also need rules of the road that will form a shared global understanding of how we will collectively confront this challenge, just as all of us instinctively understand traffic lights the same way, whether we're here in New York, or in Beijing, Moscow, or Singapore."

Importance of international cooperation

Vivian highlighted a lack of trust that AI risks are being disclosed, testing is comprehensive and credible, and cooperation can be inclusive.

Since the technology has become "an arena for strategic competition," international cooperation becomes necessary, he added.

Singapore joined the call for greater safety and security of frontier AI models, led by Finland and Norway, Vivian said.

Besides contributing to the ongoing discussions, we will also make our AI governance frameworks and testing tools available to all, he added.

"The work of building trust and laying the ground for global cooperation must begin now. Ultimately, powerful engines and good brakes and common rules do not stop us or slow us down.

They give us the assurance and the confidence so that we can move faster and more effectively for the benefit of all our peoples."

Background

In recent weeks, the Australian and United States governments have detected the unauthorised activities of AI agents on their websites.

Some AI experts also cautioned about the existential threat that AI poses to mankind, sparked by a former employee of AI company Anthropic, who told CNN that he believes AI could kill all humans by 2030.

Concerns over job disruptions and personal safety are also top-of-mind globally.