Finance ministers, central bankers, and top financiers are sounding alarms over a potent new artificial intelligence model, Claude Mythos, developed by Anthropic. Concerns are mounting that this AI could significantly compromise the security of global financial systems.
Experts warn that Mythos possesses an unprecedented capability to discover and exploit cybersecurity vulnerabilities. Canadian Finance Minister François-Philippe Champagne stated at the International Monetary Fund (IMF) meeting in Washington DC that the AI's risks are significant enough to demand the attention of all finance ministers. He highlighted the challenge of dealing with an "unknown, unknown" threat that requires robust safeguards to ensure financial system resiliency.
Financial leaders, including Barclays CEO CS Venkatakrishnan, emphasized the need to thoroughly understand the AI's exposed vulnerabilities and address them rapidly. Bank of England Governor Andrew Bailey affirmed the seriousness of the development, noting its potential to facilitate cybercrime by making it easier to detect weaknesses in core IT systems. The US Treasury has also engaged major banks to test their systems ahead of Mythos's public release.
Sources indicate other powerful AI models may emerge with fewer safeguards. James Wise, chair of the Sovereign AI unit, confirmed that Mythos represents the vanguard of powerful models capable of exposing system vulnerabilities, with his unit investing in British AI companies focused on AI security and safety.