Bank of Japan Deputy Governor Shinichi Uchida told an academic conference on October 5 that AI has become "a big positive demand shock" that has put upward pressure on the economy and prices. He then conceded that the bank has no clear answer on how large or how lasting its other effects will be.
The opening remarks at the ECONDAT 2026 Fall Meeting were a framing exercise for two days of research talks. They do not address the policy rate.
Four channels for monetary policy
Uchida said AI now features in Monetary Policy Meetings at the bank and has implications for the output gap, financial conditions and "star variables". He listed four channels.
- Demand: AI is a big positive demand shock that has put upward pressure on the economy and prices.
- Supply: it could raise productivity and capital accumulation, which might in turn affect the neutral interest rate (r*).
- Financial conditions: AI has boosted stock prices, which eases conditions. Large bond issuance by AI-related companies has pushed up long-term interest rates, which tightens them.
- Labor: it may change labor markets structurally.
Each factor, he said, pulls policy in a different direction and over a different time horizon.
What he would not claim
Uchida said the directions are understood within the traditional policy framework, but the extent and timing are not: "We don't have a clear answer yet." His tentative reading is that the demand side has come first and has made financial conditions more accommodative on balance. He added that there is a risk of correction if profits do not follow. The effects on rand on the natural unemployment rate (u) are "hard to gauge" at this juncture.
On labor, he set out both sides. AI could release workers from routine cognitive tasks and speed up innovation in fields such as R&D. It could also make some human capital obsolete, particularly skills built for intellectual labor, and widen inequality between those with technological flexibility and those without. He noted that experts disagree on how fast AI will spread. Some expect rapid diffusion, while others point to "weak-link constraints" where human processing cannot keep up.
New data, with a caveat
The first half of the speech covered the bank's own analysis. Uchida said AI and big data are easing limits on computing power and data availability, and that large language models can turn unstructured data into structured data. He cited two uses of non-traditional data: high-frequency human mobility data to judge the severity of the pandemic recession in real time, and vessel-tracking data to monitor the effect of the Middle East conflict on crude oil imports and supply chains.
He warned that the usefulness of such data might depend on the state of the economy and on how relevant it is. He also said conventional statistics may not keep up with the speed of AI adoption, and that alternative data may help, as he put it, "I hope".
Uchida said he asks himself whether the bank is making balanced decisions, neither underestimating nor overestimating AI's effects, and that the bank assesses the whole economy, including sectors less touched by AI.