Federal Reserve Backs Lighter Rules for Low-Risk Bank Artificial Intelligence
Washington, Tuesday, 7 July 2026.
Federal Reserve Governor Michelle Bowman advocated for a tiered global regulatory framework, arguing that low-risk artificial intelligence applications in banking deserve a lighter supervisory touch to encourage safe innovation.
A Collaborative Global Push for AI Standards
On Tuesday, July 7, 2026, Federal Reserve Governor and Vice Chair for Supervision Michelle Bowman delivered opening remarks addressing a newly released international framework for artificial intelligence in banking [1][2]. This address highlighted a highly coordinated global effort, following the Financial Stability Board’s (FSB) June 30, 2026 release of its first consultation report on AI-related risk management [2]. Developed with key contributions from the U.S. Federal Reserve, the Treasury Department, and the Securities and Exchange Commission (SEC), the report represents a unified regulatory front [2]. To ensure the guidelines are practical, the Federal Reserve and the FSB Secretariat initiated a public consultation on July 6, 2026, inviting industry stakeholders to provide critical feedback [1].
Advocating for a Proportionality Framework
A central pillar of Governor Bowman’s advocacy is the concept of ‘proportionality’ in supervisory oversight [1]. She emphasized that lower-risk uses of AI should receive a lighter supervisory and regulatory touch, ensuring that compliance requirements remain aligned with actual risk profiles [1][2]. Bowman noted that what works or serves as an appropriate consideration for larger, highly complex institutions deploying advanced automated systems is not suitable for smaller community banks utilizing less complex AI applications [2]. This tiered approach is designed to foster responsible technological adoption without burying smaller financial institutions under excessive regulatory burdens [2].
Designing a Tailored Regulatory Framework
The draft FSB report is structured to identify safety and soundness practices that avoid being overly prescriptive or insufficiently tailored to institutional size and risk [2]. Rather than implementing a rigid, one-size-fits-all mandate, the framework establishes material business impact and specific use-case governance as the central pillars for managing AI-related risks [1]. This flexible design allows financial institutions to scale their risk management systems in accordance with the complexity of their machine learning deployments, protecting the broader financial system from systemic vulnerabilities while leaving room for operational flexibility [1][2].
A Decade of Monitoring and Singapore’s Leadership
The Federal Reserve’s focus on machine learning is built on a deep foundation; the central bank has actively monitored bank AI usage for approximately ten years [1]. The formalization of global standards accelerated in late 2025, when the FSB Standing Committee on Supervisory and Regulatory Cooperation officially initiated the development of sound practices for AI adoption [1]. This intensive international effort was led by Hern Shin Ho of the Monetary Authority of Singapore, who guided the FSB work stream to produce the comprehensive consultative report under a highly compressed timeframe [1].
The Economic Implications and Path to Finalization
As banks of all sizes noticeably increase their deployment of AI across a variety of operational use cases, the economic stakes of these regulatory decisions are high [2]. A balanced supervisory approach is critical to ensuring that financial institutions can innovate responsibly, leveraging AI tools to optimize operations, improve credit underwriting, and enhance fraud detection [2][GPT]. By preventing regulatory overreach on low-risk applications, regulators can ensure that community banks remain competitive against larger peers, preserving diversity and credit access in the broader economy [GPT].
Next Steps and the G-20 Deliverable
The FSB is currently soliciting feedback from financial institutions, technology developers, and other stakeholders to clarify areas where the consultative document may lack precision regarding material risk mitigation [2]. Although a specific closing date for the comment period has not been finalized, the feedback gathered will directly shape the final guidelines [1][2]. The FSB plans to finalize the ‘Sound Practices for Responsible Adoption of Artificial Intelligence’ report and deliver it to the U.S. G-20 presidency by December 31, 2026 [1].