SINGAPORE — The technology that investors and governments are counting on to lift the global economy is also adding pressure that threatens growth, the head of the International Monetary Fund said, urging policymakers to stop delaying painful choices on debt.
Managing Director Kristalina Georgieva told the audience at a Wednesday event in Singapore that artificial intelligence is "rapidly becoming a key driver of countries' relative fortunes in the world economy."
But the triple forces of AI advancement, soaring energy costs and record public debt are challenging the already "underwhelming" growth this decade.
"Love it, hate it, or fear it, AI is here," Georgieva said.
Tugged in two directions
Speaking ahead of a series of IMF and World Bank annual meetings that kick off next week, Georgieva framed the global economy as being tugged in two directions at once: a "negative energy supply shock" from the war in the Gulf, now in its eighth month, and a "positive demand shock" from the AI investment boom. The combined effect, she said, is "highly uneven across the world."
On the upside, global AI investment as a share of GDP will reach, and likely exceed, the amounts that went into building the railroads, electricity grid, or telecommunications network. AI hardware and related technology products already account for more than a tenth of world goods trade, she said.
The IMF estimates that AI could add up to half a percentage point to annual world growth if done right. "Going from 3% to 3.5% over a decade — that is like adding an economy the size of ASEAN to the world economy," Georgieva said.
But the benefits are likely highly concentrated. The boom largely bypasses economies less involved in the global AI supply chain, "increasing the risk of widening economic inequality across the globe," she said.
The boom also feeds the inflation worry that has dogged policymakers from the U.S. to Europe and Asia. "The AI building boom is inflationary," she said, as are energy and food shocks, tariffs and defense spending.
Oil prices have stayed above $100 per barrel as the Middle East conflict dragged on with few signs of a diplomatic off-ramp. Retail diesel prices also rose to record highs as refining capacity squeezed energy supplies.
That inflation pressure flows straight into bond markets, as bond yields in the U.S., Germany, and Japan have surged to their highest levels in decades. Ballooning long-term private bond issuance by AI-related borrowers also competes with governments for capital, although part of the rise may reflect expectations of faster growth, Georgieva said.
The debt problem
Global public debt is near its highest level since World War II and on track to soon exceed 100% of GDP, with advanced economies the "worst offenders," Georgieva said. For 17 years, governments had "a relatively easy ride" because interest rates stayed below growth rates. "Higher interest rates now put an end to that."
The interest-to-growth differential is now "much less favourable" and "set to climb higher," she said, which means the growth needed to reduce debt ratios without fiscal effort is now "out of reach in the near term."
The strain is already visible in Europe, where spreads over German bunds are widening not just for France and Italy but for Ireland, Portugal and others that had reduced debt and deficits after the euro-area crisis.
After a string of shocks that have ballooned public debt and most countries' fiscal deficits remained above pre-pandemic averages, "fiscal space is crying out for replenishment," Georgieva said.
AI risk underpriced
Georgieva also pointed to a financial stability risk in the AI boom itself. Strong corporate earnings are driving share prices and wealth effects, she said, but "should earnings fall short, however, hyperscaler leverage and large and growing global holdings of U.S. equities could turn a disappointment into a far-reaching shock."
Citing Amara's Law, which holds that people overestimate a new technology in the short run and underestimate it in the long run, she said it is "somewhere in the transition between today's AI building boom and tomorrow's arrival of AI's benefits that we will traverse the period of maximum risk."
Georgieva said the first line of defense is regulation and supervision. "Now may be a good time for a prudently hawkish bias in many countries' monetary policy," she said.