This calls for ‘a gradual, data-dependent approach’ by central banks, says Edward Robinson
[SINGAPORE] Central banks may be able to ease monetary policy once artificial intelligence delivers widespread productivity gains, but they should keep policy appropriately restrictive for now as the AI investment boom fuels demand, said Monetary Authority of Singapore (MAS) chief economist Edward Robinson on Tuesday (Oct 20).
While AI-related investments are driving economic growth, the technology has yet to deliver significant productivity improvements across the wider economy, he said.
Speaking at the 2026 Beijing Financial Street Forum, Robinson cautioned that misjudging this transition could carry risks, calling for a gradual, data-dependent approach to monetary policy.
MAS tightened monetary policy in April and July this year amid rising inflation risks, before leaving its policy settings unchanged in October.
In its latest monetary policy statement on Oct 14, the central bank warned that domestic inflation could remain elevated if the AI-related boom generated broader economic spillovers and pushed up unit labour costs.
Still, Robinson noted that the AI boom has been particularly beneficial for Asia, with AI-driven electronics accounting for more than 70 per cent of the region’s export growth this year.
The Business Times turns 50
Five decades of milestones and moments that shaped Singapore’s success story – told through our headlines.
Explore BT50
Beyond manufacturing, Asian economies – including China – are also making progress in developing AI technologies, ranging from advanced chips to large language models.
However, Robinson, who is also MAS deputy managing director for economic policy, said growth in industries producing AI-related technologies has largely been driven by increased investment and more intensive use of existing resources, rather than improvements in efficiency.
New data centres are being built, advanced equipment is being used more intensively, and workers and machines are effectively “working harder”, rather than “working smarter”, he noted.
“AI is diffusing across the wider economy faster than the personal computer and the Internet did,” he said. “While early evidence points to efficiency gains in specific applications, it remains too soon to identify a decisive economy-wide step-up in productivity.”
Nevertheless, Robinson pointed to the potential for substantial gains over time, drawing parallels with the introduction of electricity a century ago.
The productivity benefits of electrification became substantial only after factories and workflows were redesigned around the technology, he said.
Similarly, AI could require widespread adoption, investments in skills and infrastructure, and changes to how organisations operate before its full economic benefits emerge, he added.
With sufficiently broad adoption and appropriate investment, AI could raise total factor productivity growth by up to 0.6 percentage point annually, Robinson said, pointing to estimates from the Organisation for Economic Co-operation and Development.
“This could meaningfully offset the negative effects from demographic ageing and other structural headwinds in Asia,” he noted.
