US Treasury and Federal Reserve Convene Emergency Meeting Over New AI Cybersecurity Threat
In an urgent response to emerging cybersecurity concerns, US Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell held a high-level meeting this week with chief executives from major banks. The discussion centered on the potential risks linked to Anthropic’s latest artificial intelligence model, Mythos, according to two individuals familiar with the confidential talks.
Anthropic, an AI research company, recently unveiled its powerful Mythos model. However, the organization halted a broad public release due to fears that the technology might reveal previously unknown vulnerabilities within critical digital infrastructure. According to Anthropic, Mythos possesses the ability to identify and exploit weaknesses across a wide spectrum of operating systems and web browsers, raising alarms about the potential for significant cyber threats.
Just last week, Anthropic disclosed ongoing discussions with US government officials to assess both the offensive and defensive cyber capabilities associated with Mythos. This proactive engagement underscores the serious implications of the model.
Anna Rathbun, founder and CEO of Granadilla Advisory, weighed in on the situation, highlighting the sophistication of AI’s current capabilities. She said, “AI has gotten so smart that it can actually wreak havoc on global systems like banking. This isn’t just casual use or AI making mistakes that humans can fix. This is happening at a very sophisticated enterprise level.” Rathbun expressed cautious optimism, pointing out that the developers’ decision not to distribute Mythos widely and their willingness to collaborate with regulators and banking leaders helps mitigate some concerns.
Access to Mythos will be tightly controlled, with Anthropic limiting availability to around 40 key technology companies, including major players like Microsoft and Google.
The recent meeting aimed to ensure that banks fully understand the risks posed not only by Mythos but also by similar advanced AI models. Attendees reportedly included CEOs from Citigroup, Morgan Stanley, Bank of America, Wells Fargo, and Goldman Sachs. Notably, JPMorgan CEO Jamie Dimon was unable to participate. While Goldman Sachs, Wells Fargo, and the Federal Reserve declined to comment on the meeting, other parties involved—including the Treasury, numerous banks, and Anthropic—also did not immediately respond to requests for statements.
As artificial intelligence continues to evolve at a rapid pace, the intersection of AI and cybersecurity remains a critical concern for financial institutions and regulators alike. Continued collaboration between industry leaders and government officials will be key to navigating these complex challenges and safeguarding global financial systems.