Closing the gap between Singapore's enthusiastic AI use and its lingering workplace anxiety

Singapore's numbers on AI adoption look enviable on paper. Roughly 89 per cent of professionals here now use the technology at work, a figure that routinely puts the city-state ahead of its regional peers. Yet those same workers are markedly less convinced the technology is good for their careers — only about two in three expect AI to affect their working lives positively. That mismatch between enthusiastic usage and uneasy expectations is the quiet tension running beneath Singapore's otherwise confident AI push, and it is worth unpacking before it hardens into something harder to unwind.

The observation comes from a letter to The Business Times responding to earlier reporting on worker optimism around AI. The letter points out that the gap will not close through general AI training alone. It argues for pairing instruction with supervised workplace trials: mixed-experience teams using approved tools on recurring tasks, with clear human review and protection for workers who flag failures. The suggested measure of success is practical — did AI save time without increasing errors or rework, did it improve customer outcomes, and can employees actually spot confident-but-wrong output, bias or privacy risks?

This lands at a moment when Singapore's official skilling machinery is scaling hard. The stated ambition to train 100,000 workers to apply AI is among the most aggressive national upskilling targets in the region, backed by SkillsFuture credits, AISG programmes and employer-linked pathways. But a training completion is not the same as a capability used well on the job. The distinction matters for the credibility of the whole workforce strategy: if adoption stays shallow and anxieties stay high, the gap between headline numbers and lived experience could dent confidence in the direction of travel.

For businesses in Singapore, the practical read is straightforward. The evidence the letter calls for is the kind most companies already have access to but rarely instrument — before-and-after error rates on AI-assisted tasks, customer outcome shifts, and structured channels for employees to flag unreliable output. Building those checks in early is cheaper than rebuilding trust after a high-profile failure, and it directly addresses the displacement, unfair-decision and obsolescence concerns workers keep citing.

Why it matters for Singapore: Singapore has bet heavily that its workforce, not just its infrastructure, is the draw that keeps AI investment here. A workforce that uses AI widely but feels uncertain about it is a soft spot in that pitch. Turning what is currently abstract anxiety into testable workplace evidence — with real numbers, real review loops and transparent role redesign — is how the city-state converts its headline usage rates into durable, trustworthy productivity. The 100,000-worker training goal is only the first mile; proving AI helps people do their jobs better is the rest of the journey.