Bessent and Powell issue urgent warning to bank CEOs following concerns about Anthropic model risks.
In a decisive move reflecting growing concerns over cybersecurity, Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened an urgent meeting with Wall Street executives in Washington. This gathering was prompted by apprehensions regarding the implications of Anthropic PBC’s latest artificial intelligence model, Mythos, and its potential to escalate cyber threats within the financial sector.
The meeting took place at the Treasury’s headquarters on Tuesday, where high-ranking officials aimed to ensure that major banks are cognizant of the risks associated with advanced AI technologies. Executives from systemically important banks gathered for the discussion, which coincided with their attendance at a Financial Services Forum event. These institutions are deemed critical to the stability of the global financial infrastructure, making their preparedness against emerging cyber risks a priority for regulators.
Powell’s involvement underscored the gravity of the situation, indicating that the concerns were tied to systemic risks rather than previous political tensions connected to Anthropic. The Federal Reserve’s extensive network of examiners is well-versed in the intricacies of banking operations, further highlighting the seriousness of the dialogue.
In response to the potential threats posed by Mythos, other regulatory bodies have also begun to take precautionary measures. For instance, the Bank of Canada, along with major banks and financial institutions, recently engaged in discussions regarding cybersecurity vulnerabilities associated with Anthropic’s new model. Similarly, the Bank of England has scheduled meetings with leading bank and insurance executives to assess their preparedness against such cyber threats.
Anthropic’s Mythos is characterized as a powerful AI system capable of identifying and exploiting weaknesses in major operating systems and web browsers. This capacity raises alarms among regulators, who are acutely aware of the precariousness of financial data security. In an effort to mitigate risks, Anthropic has restricted the initial rollout of Mythos to a select group of technology and financial companies, including heavyweights such as Amazon and JPMorgan Chase, under the collaborative initiative known as “Project Glasswing.” This initiative aims to safeguard critical systems before similar AI models are launched more broadly.
During testing, Anthropic’s security team revealed that the AI could potentially compromise a web browser, allowing a hacker’s setup to access sensitive information from other sites, including banking data. Such capabilities have prompted regulators and financial institutions alike to adopt a proactive stance in fortifying their defenses against an evolving threat landscape.
As the financial sector grapples with the implications of advanced AI technologies, the actions taken by Treasury and the Federal Reserve signal an urgent call to bolster cybersecurity measures and prepare for the unforeseen challenges posed by this innovative, yet potentially perilous, technological landscape.
