newsfilter
← Latest News
Federal Reserve Speeches·

Bowman, Artificial Intelligence in the Financial System

  • The Financial Stability Oversight Council convened a roundtable with public & private participants to discuss the role of AI in cybersecurity & risk management.
  • Rapid evolution of AI tools has made their risks & benefits more tangible, with examples like Anthropic's Mythos model highlighting capabilities in identifying cyber vulnerabilities.
  • AI serves as a force multiplier for the financial system, improving efficiency & speed across banks of all sizes, though smaller institutions face resource challenges.
  • Federal Reserve supervisors are ensuring banks understand inherent risks while maintaining a path for innovation & flexible implementation consistent with unique business structures.
  • Supervisory focus has shifted to identifying & remediating material financial risks that could lead to bank failure, requiring adaptable guidance for AI adoption.
  • The Fed, OCC, & FDIC amended model risk management guidance to clarify it does not apply to generative or agentic AI, reserving it for traditional models & basic applications.
  • Regulators are updating third-party risk-management guidance to reduce vagueness & ensure supervisory rules do not hinder access to new technologies.
  • Concerns regarding Anthropic's Mythos include its dual-use nature for both enhancing cybersecurity & potential malicious exploitation of weaknesses.
  • The Fed coordinates with the Treasury & other agencies to stay abreast of developments, communicate risks to banks, & incorporate industry feedback into supervisory approaches.
  • Internationally, the Fed seeks consistency in AI regulatory expectations through the Financial Stability Board to ensure a level playing field & address incompatible cross-border requirements.
  • A consultation draft report on sound AI practices, benefits, & challenges is expected to be released in the third quarter for stakeholder comment.