Federal Reserve develops new framework for AI adoption as Mythos highlights cybersecurity risks

Fed's Bowman Says Mythos Shows 'Dynamic Nature' of AI Tools

The Federal Reserve's top bank cop said regulators must consider how best to supervise new technology like Anthropic PBC's Mythos, as officials grapple with the potential risks of artificial intelligence models. "On one hand, this capability enables firms to address self-identified vulnerabilities thereby enhancing cyber security," Fed Vice Chair for Supervision Michelle Bowman said in prepared remarks on Friday. "But on the other hand, if used maliciously it could be deployed to identify and exploit weaknesses." Anthropic has limited the release of its latest AI model as the company weighs guardrails for the powerful new technology. That has led Trump administration officials to consider the potential for a new level of cyber attacks that pose a threat to financial stability. Treasury Secretary Scott Bessent and Fed Chair Jerome Powell recently met with Wall Street banks to discuss the possible risks raised by Mythos and to make sure lenders are taking precautions to defend their systems, Bloomberg reported last month. Bowman said Friday that type of meeting is "extremely beneficial" to ensuring the protection of the banking system. "Anthropic's Mythos -- an AI model that identifies cyber vulnerabilities -- highlights the dynamic nature of this technology and the rapid pace that its capability can evolve," Bowman said. The Fed's top bank cop emphasized that regulators must coordinate across government and discuss emerging technology with banks while supporting the rapid evolution of AI tools. She added that regulators are working on a report on sound practices for AI adoption and use.

Fed Prepares New Playbook to Support Bank AI Adoption | PYMNTS.com

By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions. In a speech delivered at an event hosted by the Financial Stability Oversight Council in Washington, D.C., Bowman said regulators must ensure safety and soundness without hindering innovation. The Mythos AI model's ability to detect cyber vulnerabilities can be used for enhancing cybersecurity or for malicious purposes, Bowman said. "As we learn more about this tool and others to be released in the coming weeks and months, we will continue to consider effective supervisory approaches for these and other emerging capabilities," Bowman said. Bowman said that regulators' supervision of emerging technologies requires keeping up with new developments, coordinating efforts across government, communicating developments and risks to supervised institutions, getting feedback from industry, and regularly refining supervisory approach and response. "As we work to support innovation, it is necessary to determine whether our framework is appropriate," Bowman said. "Have we established reasonable and effective supervisory expectations? Are bankers comfortable discussing emerging risks and new technologies with supervisory teams? Have we successfully implemented a pro-innovation mindset that allows responsible innovation and AI adoption to occur within the banking system?" Bowman said that the Financial Stability Board's Standing Committee on Supervisory and Regulatory Cooperation, of which she is the chair, is working with the Treasury Department and the Securities and Exchange Commission (SEC) on a report that will cover sound practices for AI adoption, use and innovation. A draft is set to be released in the third quarter. "The implications of AI extend far beyond the banking and financial systems," Bowman said. It was reported April 7 that Anthropic was allowing select partners to gain early access to Claude Mythos Preview, a model positioned for defensive cybersecurity work, so that they could identify vulnerabilities and strengthen systems before threats can be exploited. Anthropic said at the time it was also discussing the model and its "offensive and defensive cyber capabilities" with government officials.