BIS: AI Is Blurring the Reference Indicators for Central Bank Rate Decisions

Cailian Press, October 8 (Editor: Li Ying) — The latest research from the Bank for International Settlements (BIS) shows that the impact of AI on productivity, energy and climate is blurring the economic indicators that central banks generally reference when setting interest rates.

In a research report released on Thursday local time, the BIS said that AI is changing potential output through channels such as productivity, climate resilience and energy constraints, making the key indicators on which central banks rely when setting interest rates increasingly difficult to assess.

Potential output is the level of output an economy can achieve when resources such as labor and capital are sustainably and fully utilized and inflation will not continue to rise. This indicator cannot be directly observed, and institutions such as central banks generally estimate it using historical data and economic models.

When setting interest rates, central banks compare actual output with potential output to judge whether capacity is running hot or cold: actual output above potential output means resources are tightening, inflationary pressure is rising and the economy may be overheating, which tends to favor raising rates; the reverse means capacity is idle and there is room to cut rates.

Rising energy costs and grid bottlenecks raise firms' production costs and limit the connection of new capacity, thereby pushing down the economy's potential output.

Once potential output changes, central banks' judgments may deviate. For example, if AI has already raised potential output but the central bank fails to notice, it may misjudge normal growth as overheating and raise rates too early; if energy constraints drag down potential output, the central bank may overestimate the economy's slack, causing inflation to pick up.

The report argues that this makes it harder for central banks to distinguish whether inflationary pressure comes from cyclical demand overheating or from structural changes in production capacity, and thus harder to assess potential output.

The former can usually be cooled by raising rates, while the latter may not be, and raising rates may also inadvertently harm growth.

Financial risks loom

This concern is consistent with warnings the BIS has repeatedly issued. The institution has repeatedly stressed that a crash after the AI boom could cause widespread economic damage.

The BIS's "Annual Economic Report 2026," released on June 28, compared the current AI investment boom with the 19th-century canal and railway mania, the electrification boom of the 1920s and the internet bubble of the 1990s, arguing that its scale and speed are similar to these historical precedents.

In September, BIS General Manager Pablo Hernández de Cos said in a speech that valuations of AI companies are built on "ambitious expectations" and have triggered an "arms race" among these companies.

Sovereign bonds of many countries continue to "bleed" as Google and Wolfspeed move and surge | Tonight's highlights

Collapse; Fed's Waller: Further rate hikes still necessary, but no need for continuous action

Collapse; Musk appears to publicly lash out at India's richest man, prompting the Indian government to issue a statement rebutting it