Humans Must Stay in Control of AI Defence Decisions: Chan Chun Sing

Defence Minister Chan Chun Sing has drawn a clear line on military AI: automation can speed up targeting and intelligence analysis, but a human must always remain in the decision loop. Speaking at The Straits Times' inaugural Forum dialogue on June 24, Chan warned that handing combat decisions entirely to algorithms would amount to a "dereliction of duty."

Chan acknowledged that AI is already reshaping military operations — helping the Singapore Armed Forces (SAF) interpret sensor data, acquire targets, and recommend weapon systems faster than humans alone could manage. But he stressed that this acceleration must never come at the cost of understanding. "It cannot be so autonomous, whereby you just take decisions, and you don't know what you are hitting," he said. The message comes as global defence forces increasingly experiment with autonomous drones and AI-powered command systems, raising ethical and operational questions that Singapore is now addressing pre-emptively.

The minister's stance aligns with a broader push across the SAF to modernise while keeping human judgment central. Singapore maintains a defence spending policy of up to 6 per cent of GDP, though actual spending has hovered around 3 per cent in recent years. The SAF draws on a core of 50,000 high-readiness personnel supported by 250,000 Operationally Ready NSmen. Chan also noted that women now make up more than 10 per cent of regular servicemen and women, with plans to expand the SAF Volunteer Corps to accept a wider pool of applicants, including first-generation permanent residents.

Why it matters for Singapore: Chan's remarks are the clearest articulation yet of where Singapore draws the line on autonomous weapons. As a small state that relies on deterrence through readiness rather than mass, Singapore cannot afford either to fall behind in AI adoption or to cede control to black-box systems. The "human-in-the-loop" doctrine reassures Singaporeans that the SAF's deterrent power rests on trained judgment, not unaccountable algorithms. It also positions Singapore's voice in global AI governance debates — particularly relevant as the Pentagon and other major militaries push for broader AI roles in targeting.