Treasury, Fed Urge Goldman Sachs to Strengthen Security Against Anthropic’s Mythos AI Risks
Treasury Secretary Scott Bessent and Fed Chair Jerome Powell summoned Goldman Sachs CEO David Solomon and other bank leaders to warn about cybersecurity risks from Anthropic’s Mythos AI model. Mythos identified thousands of zero-day vulnerabilities in major operating systems and browsers, driving banks to enhance AI-driven security defenses.
1. Treasury and Fed Convene Bank CEOs
On April 7, Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell called CEOs of systemically important banks, including Goldman Sachs’ David Solomon, to a meeting in Washington. The session focused on the potential cybersecurity threats posed by advanced AI models and the need for coordinated responses across the financial sector.
2. Mythos AI Model Exposes Systemic Vulnerabilities
Anthropic’s new Mythos AI demonstrated the ability to uncover thousands of zero-day vulnerabilities across dominant operating systems and browsers, including a 27-year-old flaw in OpenBSD and an undetected video-library issue after five million automated tests. These findings highlight AI’s capacity to both reveal and exploit critical security gaps in financial infrastructure.
3. Banks to Enhance AI-Driven Cybersecurity
Following the warnings, Goldman Sachs and its peers are expected to allocate additional resources and adopt AI-powered defenses under initiatives like Anthropic’s Project Glasswing, which offers up to $100 million in usage credits to partners. Banks will need to integrate these tools swiftly to safeguard client data and maintain regulatory compliance.