AI is making it harder for central banks to determine an economy’s potential growth rate.

Mumbai: AI is making it harder for central banks to determine an economy’s potential growth rate, as technological change increases uncertainty around productivity, potential output and the natural rate of interest, BIS general manager Pablo Hernández de Cos said.“AI is reshaping demand, supply and financial markets simultaneously, making the economy harder to read and increasing uncertainty around the transmission of monetary policy,” Hernández de Cos said at the Global Fintech Fest in Mumbai on Thursday.While the objectives of monetary policy remain unchanged, AI-driven structural changes could make some of the key variables that guide policy harder to estimate, he said. Potential output and the natural rate of interest are already subject to significant uncertainty, and rapid changes driven by AI could widen measurement errors and increase discrepancies between actual and estimated output gaps.The longer-term economic impact of AI is also uncertain, Hernández de Cos said, pointing to three scenarios outlined in the BIS Annual Economic Report. In one, AI produces a bounded productivity boost that permanently raises the growth trend. In another, AI generates self-reinforcing, transformative growth as the technology improves itself. A third scenario, the “demand bottleneck”, sees automation diverting income from labour towards further AI investment, eventually weakening consumer spending and causing productivity and output growth to stall.These scenarios could have very different implications for the natural rate of interest and inflation, he said. Stronger productivity growth could raise r-star, while a demand bottleneck could eventually push it below pre-AI levels and become disinflationary.“The takeaway from this analysis is that there is considerable uncertainty about the long-term effects of AI,” Hernández de Cos said. The eventual outcome would depend on how much AI progress relies on consumer demand, how competition affects profit margins in the AI sector and how quickly AI products and infrastructure depreciate.Hernández de Cos flagged financial-stability risks from the AI investment boom, with concentrated equity valuations, capex outpacing cash flows and rising reliance on debt/private credit. Opaque “circular financing” between chipmakers, hyperscalers and AI firms could amplify losses if AI returns disappoint, turning the investment boom into a bust.AI could also deliver major productivity gains, with studies showing 10-65% task-level gains and 20-50% time savings; median estimates suggest a 0.5 percentage-point annual boost to total factor productivity. But benefits will be uneven: AI may augment skilled work while displacing routine cognitive jobs, with informal workers in developing economies particularly vulnerable. Early displacement is emerging in customer support, programming and administration, making reskilling essential.

Advanced economies are better placed to capture near-term gains, though preparedness varies sharply among emerging markets. India has an opportunity to close the gap by building on its digital public infrastructure. Rather than replicate full AI ecosystems, countries should target comparative advantages across chips, cloud, data, models and applications, including local-language models and domain-specific data.For central banks, AI raises the premium on measurement, judgement, policy agility and international cooperation. Its economic impact will ultimately depend on policy choices, including investment in digital infrastructure and skills, competition, data governance and sound institutions.