US Banks Warned of Cyber Risks from Anthropic's Mythos AI Model

Severity: High (Score: 67.8)

Sources: Cnbc, Theguardian, Fstech, Afr, Thebanker

Summary

US Treasury Secretary Scott Bessent and Federal Reserve Chairman Jerome Powell convened a meeting with major U.S. bank CEOs to discuss the cybersecurity threats posed by Anthropic's newly launched AI model, Claude Mythos. The meeting was prompted by concerns that the model could be exploited by hackers, as it has already identified thousands of software vulnerabilities, including zero-day flaws. Attendees included executives from Bank of America, Citigroup, Goldman Sachs, Morgan Stanley, and Wells Fargo, while JPMorgan's Jamie Dimon was unable to attend. The banks are considered systemically important, meaning disruptions could impact financial stability. Anthropic has limited the model's rollout to select partners like Amazon, Apple, and Microsoft due to these risks. The meeting reflects heightened anxiety over AI's potential to exacerbate cybersecurity threats, with the Trump administration actively engaging in discussions about the implications of such technology. The situation is further complicated by ongoing legal disputes regarding Anthropic's designation as a supply chain risk by the Department of Defense. Key Points: • US Treasury and Federal Reserve convened bank CEOs to discuss Anthropic's Mythos AI risks. • Mythos has identified thousands of vulnerabilities, raising concerns over potential hacker exploitation. • The model's rollout is limited to select partners due to fears of misuse in cyberattacks.

Key Entities

• Zero-day Exploit (attack_type)
• Project Glasswing (campaign)
• Iran (country)
• Financial (industry)
• Claude Mythos Preview (platform)