BIS Submits Report on AI Use for Policy Purposes to G20 Finance Ministers and Central Bank Governors

The Bank for International Settlements published a report on October 10, 2025 examining the use of artificial intelligence for policy purposes and formally submitted it to G20 Finance Ministers and Central Bank Governors. The report focuses on how AI tools are being applied within policy and regulatory functions at central banks and financial regulators, representing one of the most senior-level multilateral reviews of AI in financial governance to date. Although no specific binding obligations or numerical thresholds are attached to the report, its submission to the G20 places AI governance squarely on the agenda of the world's most consequential economic coordination forum.

The report reflects an accelerating trend of international regulatory bodies attempting to establish shared frameworks for AI use before domestic supervisory regimes fragment into incompatible standards. Central banks and financial regulators across G20 member states have been independently piloting AI tools for functions such as market surveillance, systemic risk monitoring, and regulatory reporting analysis. The BIS report addresses the governance gaps and consistency challenges that arise when these efforts proceed without common principles, signaling that multilateral coordination bodies view AI policy use as a systemic-level concern rather than a purely operational one.

Enterprise compliance teams at financial institutions operating across G20 jurisdictions should treat this report as an early indicator of supervisory expectations that are likely to materialize in domestic guidance over the next one to two years. Firms should begin mapping their own use of AI in any process that touches regulatory reporting, risk modeling, or policy interpretation, as these are the areas most likely to draw scrutiny once member jurisdictions translate BIS guidance into national frameworks. Compliance officers should engage with their prudential supervisors to understand whether existing model risk management policies, such as those governed by SR 11-7 in the United States or equivalent frameworks elsewhere, will be extended to cover policy-adjacent AI applications, and should document governance structures for those tools now rather than retroactively.