Artificial Intelligence in Europe: does advanced AI have a glass ceiling?

Thank you for the opportunity to present on the Financial Stability Board's (FSB) consultation report addressing Responsible Adoption of Artificial Intelligence (AI) in the financial sector. Federal Reserve staff and FSB Secretariat personnel are eager to incorporate feedback in finalizing the report, scheduled for release later this year as a U.S. G-20 contribution.

First, I extend gratitude to Hern Shin Ho from Singapore's Monetary Authority for their leadership in this FSB work stream, completed in a tight timeline. Recognition also extends to FSB Secretariat staff who contributed to this report's success. Additionally, I wish to thank Treasury and SEC colleagues for their collaboration in producing this document.

As FSB Chair of the Standing Committee on Supervisory and Regulatory Cooperation, I initiated work on sound practices for AI adoption in financial institutions last year. This report reflects ongoing AI advancements and broader financial institution usage of the technology. It seeks public commentary on AI's benefits, associated risks, and current management practices.

The Federal Reserve has monitored bank AI usage for over a decade, noting a recent surge in adoption across all institution sizes. Our focus remains on supporting responsible innovation through AI adoption. We have provided illustrative governance examples and case studies within the FSB report to guide responsible AI deployment.

Key to managing AI risks is understanding specific use cases. The report offers a range of examples and case studies to illustrate appropriate governance and controls. Financial institutions must be explicit about AI usage and its materiality to business operations or legal obligations. This will inform the level of required governance and controls.

The report emphasizes that lower-risk AI applications warrant lighter oversight, while higher-risk uses require more rigorous supervision and regulation. Proportionality is a critical factor in applying the appropriate level of oversight to financial institutions regardless of size or complexity. The guidance provided aims to support innovation while mitigating risks across the financial system.

In conclusion, this report marks a crucial beginning in addressing AI in finance. We invite feedback on the report's balance, particularly concerning proportionality. Addressing institutional size, complexity, and risk profile is paramount. Furthermore, we need clarification on adequately addressing material risks and additional guidance for managing AI-related risks effectively.

Your input during today's session and the upcoming public comment period will aid in formulating the final report to be presented to the G-20 later this year.