AI enters the central bank toolkit: Uchida's four channels, Basel's leverage warning, BIS's circular deals

Serguey Asael Shinder05 October 2026, 07:242 min read

Within a few days, three central bank voices treated artificial intelligence not as a technology to adopt but as a variable in their own frameworks.

Monetary policy. Bank of Japan Deputy Governor Shinichi Uchida set out on 5 October how AI enters rate-setting: as a positive demand shock, as a possible supply-side change that could move the neutral rate, through financial conditions eased by higher stock prices and tightened by heavy AI-related bond issuance, and through structural labour-market change. His tentative reading was that the demand side came first and eased conditions on balance, with a risk of correction if profits do not follow.

Bank supervision. The Basel Committee said on 1 October that the financial footprint of the AI ecosystem is expanding rapidly, with more leverage and increasingly interconnected financing, and that frontier AI could amplify operational vulnerabilities through cyber attacks and correlated dependencies. It will review its operational-risk loss categories with AI in mind.

The financing web. A BIS Bulletin published the same day measured what "interconnected financing" means: between 2021 and 2025, 28.7% of AI firms' investment deals by value went into other AI firms, and 55.2% of incoming investment in AI firms came from other AI firms. Of AI-to-AI deals, 46.4% by value also involved a commercial supply relationship between investor and target. The authors say such circular relationships reflect real economic needs, such as securing critical inputs, but carry macroeconomic risks and increase opacity.

The common thread. All three describe the same features from different positions. Uchida sees easier financial conditions driven by AI equity values and a risk if profits disappoint. Basel sees leverage and interlinked financing in the banks' exposure. The BIS bulletin shows why the two connect: when investors are also suppliers and customers of the firms they fund, valuations and revenues can reinforce each other on the way up and on the way down.

What it means

None of this is a policy change, and each source is careful to say how little is yet measured. But the vocabulary has shifted: AI now appears in discussions of the output gap, r*, operational-risk taxonomies and investment networks. The practical consequence is that a correction in AI valuations would no longer be treated only as a market event; central banks are already describing the channels through which it would reach policy and bank balance sheets.