Governance

SR 11-7

Federal Reserve guidance on model risk management applicable to AI/ML in banking.

Definition

SR 11-7 is the Federal Reserve's Supervisory Guidance on Model Risk Management, issued in 2011. While predating modern AI/ML, it remains the foundational regulatory framework for managing model risk in US financial institutions and has been interpreted to apply to AI systems.

Key principles:

  • Model definition: Any quantitative method that processes input data into quantitative estimates—including AI/ML
  • Three lines of defense: Model developers, independent validation, and internal audit
  • Effective challenge: Critical analysis by qualified, independent parties
  • Documentation: Comprehensive records of model purpose, design, limitations, and testing

Banks must demonstrate that AI models meet SR 11-7 standards, creating significant compliance requirements for AI adoption in financial services.

Sources

  • Federal Reserve SR 11-7
  • OCC 2011-12