In a development that underscores the growing intersection of artificial intelligence and financial stability, Anthropic has agreed to brief the Financial Stability Board (FSB) on the cybersecurity implications of its Claude Mythos AI model. The move comes amid mounting alarm over the potential for advanced AI systems to uncover and exploit software vulnerabilities at scale, posing systemic risks to global finance.
Anthropic’s Mythos model, designed to detect previously unknown flaws in IT systems, has already demonstrated capabilities that alarm cybersecurity experts. The UK’s AI Security Institute (AISI) reported that Mythos completed a previously unsolved cybersecurity test—dubbed “cooling tower”—in three out of ten attempts, marking a notable leap in autonomous cyber task performance (theguardian.com).
Rather than releasing Mythos publicly, Anthropic has restricted access to a select group of technology firms and banks—including Apple and JPMorgan—to help them identify and remediate vulnerabilities (theguardian.com). The FSB, chaired by the governor of the Bank of England, will now receive a briefing on the model’s capabilities and associated risks (theguardian.com).
The International Monetary Fund (IMF) has also sounded the alarm, warning that AI models like Mythos could enable cyberattacks capable of disrupting payments, straining liquidity and solvency, and threatening financial stability. The IMF urged immediate investment in cyber defenses, enhanced monitoring, stress testing, and coordinated oversight to mitigate contagion risks (asset-management-news.com).
This development highlights a critical inflection point: AI systems that can autonomously identify vulnerabilities may offer defensive benefits—but also introduce new attack vectors. As Federal Reserve Vice Chair Michelle Bowman noted, tools like Mythos “show the dynamic nature of this technology and how quickly its capabilities can develop,” raising urgent questions about oversight and governance (cryptonews.net).
With regulators and financial institutions now grappling with the dual-use nature of frontier AI, the briefing to the FSB marks a pivotal step toward coordinated global oversight. The outcome could shape how AI is deployed—and regulated—in critical financial infrastructure going forward.