US Treasury Summons Bank Chiefs Over AI Cybersecurity Risks from Anthropic's Claude Mythos
US Treasury Secretary Scott Bessent summoned the chief executives of major American banks to an urgent meeting in Washington this week, following reports of unprecedented cybersecurity risks posed by Anthropic's latest artificial intelligence model, Claude Mythos. Federal Reserve Chair Jerome Powell was among the senior officials who attended the gathering at Treasury headquarters, highlighting the severity of regulatory concerns.
Unprecedented Vulnerabilities Exposed
Anthropic, the AI startup behind Claude Mythos, recently disclosed that its model has identified thousands of previously unknown vulnerabilities in widely used software and popular applications. In a blog post published earlier this month, the company warned that AI models have now surpassed "all but the most skilled humans at finding and exploiting software vulnerabilities," adding that "the fallout – for economies, public safety, and national security – could be severe."
The company revealed that some of the vulnerabilities discovered by Mythos date back up to 27 years, none of which had been detected by their original developers or technology monitors before the AI model identified them. This revelation has raised alarms about potential exploitation by malicious actors who could use such tools to compromise passwords, crack encryption, and access sensitive data.
Banking Sector on High Alert
The Washington meeting focused specifically on heads of systemically important banks – institutions whose potential disruption or collapse could threaten overall financial stability. Attendees included Goldman Sachs CEO David Solomon, Bank of America's Brian Moynihan, Citigroup's Jane Fraser, Morgan Stanley's Ted Pick, and Wells Fargo's Charlie Scharf, according to Bloomberg's initial report. JP Morgan's Jamie Dimon was invited but unable to attend.
In his annual shareholder letter published this week, Dimon emphasized that cybersecurity "remains one of our biggest risks" and warned that "AI will almost surely make this risk worse." This sentiment reflects growing apprehension within the financial sector about the dual-edged nature of advanced AI capabilities.
Restricted Release and Legal Challenges
In response to the security concerns, Anthropic has taken the unprecedented step of limiting Claude Mythos's release to a select group of businesses, including technology giants Amazon, Apple, and Microsoft. Networking companies Cisco and Broadcom have also gained access, along with the Linux Foundation, which promotes the open-source Linux operating system.
This marks the first time Anthropic has restricted product availability, underscoring the model's potential risks. The development comes weeks after the US government designated Anthropic as a supply chain risk, allegations the company is currently contesting in court.
Regulatory Response and Industry Implications
The emergency meeting was reportedly convened while bank executives were already in Washington for a lobby group gathering, allowing for immediate coordination between financial regulators and industry leaders. The Federal Reserve, Anthropic, and the involved banks declined Bloomberg's requests for comment, while the Treasury Department did not respond to inquiries.
This incident highlights the evolving challenges regulators face in balancing technological innovation with security imperatives, particularly as AI capabilities advance at an accelerating pace. The banking sector's heightened vulnerability to cyber threats makes this development particularly significant for financial stability and consumer protection.
