The Financial Stability Board (FSB) recently called upon Anthropic to conduct a closed briefing for G20 economic regulators, spotlighting the potential risks posed by the AI model Claude Mythos. This initiative highlights growing concerns among central banks and financial authorities regarding the impact of advanced AI technologies on global financial stability.
Chaired by Bank of England Governor Andrew Bailey, the meeting will bring together representatives from major economies, including the United States, China, Japan, Germany, and France. The urgency of this gathering arises from findings that indicate new AI models like Claude Mythos could reveal critical vulnerabilities within banking infrastructure, potentially leading to systemic risks.
In early April, Anthropic launched Claude Mythos Preview, an AI model primarily focused on cybersecurity. This model has autonomously identified thousands of zero-day vulnerabilities across various software platforms. The implications of these discoveries are significant, raising concerns about their potential effects on economies, public safety, and national security.
In response to these serious vulnerabilities, Anthropic has initiated Project Glasswing, aimed at strengthening the protection of essential infrastructure. The company has committed over $100 million in grants and loans to support this initiative, demonstrating its dedication to mitigating the risks associated with its AI technologies.
The concern over AI-driven threats has reached a critical level, prompting U.S. regulators to act swiftly. An emergency meeting among leading Wall Street banks was convened, attended by influential figures such as U.S. Treasury Secretary Scott Bessent and former Federal Reserve Chair Jerome Powell. This meeting was prompted by fears of a new wave of AI-driven cyberattacks that could exploit the vulnerabilities identified by Claude Mythos.
As G20 regulators prepare for their closed briefing with Anthropic, discussions will likely focus on managing the emerging risks associated with AI models like Mythos. The potential for faster and more destructive AI-driven attacks poses a pressing challenge that requires immediate attention from financial authorities worldwide.
The outcome of these discussions could influence regulatory frameworks designed to protect financial systems from the evolving threats posed by artificial intelligence. With AI technology advancing rapidly, a coordinated approach to managing its risks is more critical than ever.
