Singapore Cuts Scam Losses to S$913M, but Shanmugam Warns AI Is the New Frontline

Singapore has spent the better part of a decade building one of the world's most aggressive scam-fighting apparatus, and the numbers finally turned last year. Reported scam cases and the money lost both fell for the first time. Yet the man leading the home front is refusing to call it a win — because the technology powering the next wave of fraud is moving faster than any enforcement playbook.

Speaking on Aug 18 at the Singapore Police Force and Ministry of Home Affairs scholarship award ceremony, Senior Minister K. Shanmugam laid out the mixed picture. Total scam losses dropped to S$913.1 million in 2025 from a record S$1.1 billion in 2024, and reported cases fell more than 27 per cent to 37,308. But AI has already delivered a painful counter-argument: one victim lost S$4.9 million to a deepfake after believing he was speaking to Singapore government officials. Shanmugam also noted that wrongful use of AI was involved in the self-radicalisation of two youths, one of whom used a chatbot to pledge allegiance to ISIS while the other drew on the technology to research ammunition and 3D-printed firearms. Both were detained under the Internal Security Act.

Singapore's answer is to meet AI with AI. The SPF has stood up its Cyber Command, whose Digital Disruption Centre hunts and removes online threats, while the Central Narcotics Bureau uses AI to streamline investigative work and the Immigration and Checkpoints Authority deploys facial recognition at the border. On the legislative side, Shanmugam pointed to a legal framework he described as "peculiar to Singapore" — laws such as the Maintenance of Religious Harmony Act, the Racial Harmony Act and the Foreign Influence (Countermeasures) Act — arguing that most countries facing these threats cannot move as quickly.

The pattern here is instructive for how a small, heavily digitalised state positions itself against AI-enabled harm. Singapore is not just building harder walls; it is treating the legal system itself as an enforcement tool that can be updated faster than the technology evolves. That is the same philosophy driving the government's runaway AI investment — the country that deploys AI most aggressively is also the one designing the guardrails around it. The challenge is that the threat surface keeps widening, from financial scams to borders to the quiet radicalisation of teenagers in their own bedrooms.

Why it matters for Singapore: Shanmugam's remarks underscore a double-edged reality for the local AI landscape. Singapore wants to be a global AI hub, drawing capital and talent into the sector, but that ambition rests on public trust — and nothing erodes trust faster than a deepfake that empties a victim's bank account. The S$4.9 million deepfake case and the AI-linked radicalisation of minors are precisely the failure modes that could invite public pushback against AI adoption. Expect the government to keep pairing every AI push with visible enforcement muscle, from Cyber Command to tightened laws, as it tries to hold the line between being a leader in AI and a leader in protecting citizens from it.