Analysis frame
Primary-source evidence
Separate near-term investment demand from later productivity supply, financial conditions and labor adjustment.
- Workers whose cognitive tasks may change
- Consumers, investors and firms exposed to AI-related price and rate effects
- The speech gives no causal estimate of AI's net inflation or employment effect
- The timing and distribution of productivity gains remain uncertain
- High investment may ease equity financing while AI-related debt raises longer-term borrowing costs
- Alternative data may improve policy observation while introducing context-dependent blind spots
Four channels, not one prophecy
The speech identifies a demand shock, potential supply gains, opposing effects on financial conditions and labor-market change. A single headline about productivity misses the different mechanisms.
The deputy governor explicitly says that the magnitude and time horizon remain unclear. The present tentative judgment is that demand has arrived before supply.
A better data feed is not a crystal ball
The bank has used mobility and vessel-tracking data to study shocks, and AI can turn unstructured material into analyzable data. But its usefulness depends on the economic state and relevance of the dataset.
What matters next is evidence on productivity, wages, prices and returns together. Neither a model benchmark nor a stock index can substitute for the distribution of real gains.
Go to the source
Read the evidence behind this analysis. External links open in a new tab.
Bank of Japan — AI, big data and monetary policy Bank of Japan — full remarks PDF Bank of Japan — conference program


