United Kingdom briefed on AI-found cyber vulnerabilities | JQJO

LONDON, England – Anthropic is preparing to brief global financial regulators on cybersecurity vulnerabilities identified by its unreleased AI system, Mythos Preview, according to a Financial Times report cited by Reuters on Monday. The specialized cybersecurity-focused model is designed to identify weaknesses in software, internet infrastructure, and web browsers, including longstanding issues that conventional scanning tools may miss. The planned meeting, requested by Bank of England Governor Andrew Bailey in his capacity as chair of the Financial Stability Board (FSB), is expected to involve finance ministries, central banks, and regulators focused on AI-driven cyber risks to the global banking system and critical digital infrastructure.

LONDON, England – The briefing will outline how advanced AI systems could rapidly expose vulnerabilities across financial networks and widely used digital infrastructure at a pace that may challenge existing defenses, the report said. The FSB, created after the 2008 global financial crisis to monitor systemic risks, has previously warned that cyberattacks pose a major threat to financial stability, and regulators are increasingly examining how frontier AI models might transform both offensive and defensive cybersecurity operations. Reuters said it could not independently verify the Financial Times report and noted that Anthropic and the FSB did not immediately respond to requests for comment, while Mythos Preview remains unreleased and its detailed technical specifications undisclosed.