US summons bank bosses over cyber risks from Anthropic’s latest AI model
The Guardian • • business
Key Points:
- US Treasury Secretary Scott Bessent convened a meeting with top American bank CEOs in Washington to discuss cybersecurity risks associated with Anthropic's new AI model, Claude Mythos.
- Federal Reserve Chair Jerome Powell and leaders from major banks including Goldman Sachs, Bank of America, Citigroup, Morgan Stanley, and Wells Fargo attended the meeting; JP Morgan's Jamie Dimon was invited but did not attend.
- Anthropic revealed that its Mythos AI model has identified thousands of software vulnerabilities, some up to 27 years old, raising concerns about potential exploitation by hackers and prompting limited release to select companies like Amazon, Apple, Microsoft, Cisco, Broadcom, and the Linux Foundation.
- The meeting highlights growing worries that advanced AI tools could be used to compromise cybersecurity by cracking passwords and encryption, with Dimon previously warning that AI could exacerbate cybersecurity risks.
- This discussion follows the US government's recent designation of Anthropic as a supply chain risk, a classification the company is currently contesting in court.
