What Cyber Flaws Did Mythos AI Expose?
Anthropic is preparing to brief the Financial Stability Board on cybersecurity vulnerabilities in the global financial system exposed by its unreleased Mythos AI model, the Financial Times reported on Monday. The briefing was requested by Bank of England Governor Andrew Bailey, who chairs the FSB, and will be delivered to G20 finance ministries and central banks.
A Model Too Dangerous to Release
Anthropic announced Claude Mythos Preview in April as a specialized AI system designed to detect long-standing vulnerabilities in software, browsers, and critical infrastructure. The company has described it as capable of identifying exploitable flaws in "every major operating system and web browser," succeeding in developing working exploits on its first attempt in more than 83 percent of internal tests.
The model has not been publicly released. Instead, Anthropic has distributed access to roughly 40 trusted organizations — including Amazon, Apple, Google, and JPMorgan — through a program called Project Glasswing, enabling coordinated fixes before public disclosure. The New York Times reported that Anthropic deemed the system too potent for general release due to the potential for hackers to leverage it to exploit network vulnerabilities.
Regulators Sound the Alarm
The FSB briefing comes days after UK regulators issued a joint statement on frontier AI cyber risks. On May 14, HM Treasury, the Bank of England, and the Financial Conduct Authority warned that "the cyber capabilities of current frontier AI models are already exceeding what a skilled practitioner could achieve, and at a significantly higher speed, greater scale, and lower cost". The statement urged regulated firms and financial market infrastructures to strengthen defences, improve vulnerability management, and consider deploying AI-enabled protections.
Reuters reported that in the United States, Federal Reserve Chair Jerome Powell and Treasury Secretary Scott Bessent convened an urgent meeting with major bank CEOs in April to discuss the cyber risks linked to Mythos.
Systemic Risk Concerns
The planned FSB briefing marks a turning point in how frontier AI capabilities are being treated — not merely as a technology policy matter, but as a potential source of systemic financial risk. Bailey's decision to bring Anthropic before the board underscores a growing consensus among global financial authorities that models capable of both finding and weaponizing software flaws demand coordinated oversight at the highest levels.
