AI in action · Markets
MAS Economist Says AI Has Yet to Deliver Broad Productivity Gains and Warns Expectations Could Inflate Asset Prices
AI has yet to deliver widespread productivity gains, Singapore central bank chief economist Edward Robinson said in Beijing on October 20, 2026. Much of the growth in AI-producing industries reflects investment, not productivity, Fintech News Singapore reported. He warned that expectations of future gains could inflate asset prices and encourage excessive borrowing, and said monetary policy should follow actual conditions, not hoped-for ones.
Central bankers usually discuss artificial intelligence as a tool or a risk to supervise. Edward Robinson, deputy managing director for economic policy and chief economist of the Monetary Authority of Singapore, discussed it as a macroeconomic claim that has not yet been proven. Speaking at the 2026 Beijing Financial Street Forum on October 20, he said businesses are adopting AI rapidly but that the technology has yet to deliver widespread productivity gains, according to Fintech News Singapore's account of the speech. Much of the growth in AI-producing industries, he said, reflects higher investment and greater use of equipment rather than sustained improvements in productivity.
The distinction matters because the two look the same in the short run. Building data centres and buying chips raises output and employment in the industries that supply them, which shows up in GDP and in Asian export figures; the same article noted that AI-driven electronics account for more than 70% of Asia's export growth this year. Productivity growth, by contrast, means the rest of the economy producing more with the same inputs, and Robinson argued that broader gains would require businesses to adapt their operations and invest in skills and infrastructure. The OECD has estimated that AI could add up to 0.6 percentage points a year to productivity growth with widespread adoption, a large number, but a conditional one.
His warning was about the gap between the two. Expectations of future AI gains could inflate asset prices and encourage excessive borrowing, he said, and a reassessment could bring falling valuations and tighter credit conditions. For policy, he drew a line that many investors would rather central banks did not: monetary policy should be based on actual economic conditions, not on anticipated productivity gains, and may need to stay restrictive while investment-driven demand dominates. In other words, an AI boom that raises spending today without yet raising supply is inflationary, and a central bank should treat it that way. He also called for safeguards against excessive borrowing.
The speech fits a pattern in this week's official commentary. The Bank of Japan's Financial System Report, published three days later, added a fall in AI profit expectations to its stress test, and IBM's chief executive told analysts some corporate AI spending is "completely out of control." None of them predicted a crash. Each, from a different seat, made the same observation: the money is arriving before the measurable returns, and markets are pricing the returns.
A caveat on sourcing: Fintech News Singapore paraphrased the speech and published no direct quotations, and MAS had not posted a transcript at the time of writing, so the wording above is the outlet's summary of Robinson's argument rather than his own.
Sources: Fintech News Singapore — MAS warns AI expectations could inflate asset prices (Oct 20, 2026); Bank of Japan — Financial System Report (October 2026), Oct 23, 2026