Finance ministers and top banking leaders have raised serious concerns about a new artificial intelligence model that could pose risks to global financial systems.
The model, Claude Mythos, developed by Anthropic, has triggered urgent discussions after demonstrating its ability to detect vulnerabilities in major operating systems.
Experts warn that the model could identify and exploit cybersecurity weaknesses at a highly advanced level. However, some specialists believe further testing is necessary to fully understand its real-world capabilities.
François-Philippe Champagne confirmed that the issue received significant attention during meetings at the International Monetary Fund in Washington DC.
He described the challenge as unpredictable, comparing it to unknown risks that are harder to control than physical threats. He emphasized the need for stronger safeguards to ensure the resilience of financial systems.
Concerns surrounding Mythos appear to exceed those linked to earlier AI models. Still, some cybersecurity experts question whether the fears may be overstated, as the model has not yet undergone widespread independent testing.
The AI Security Institute has conducted one of the first independent evaluations. Its findings show that Mythos can exploit systems with weak security, but it may not represent a major leap beyond its predecessor, Claude Opus 4.
Researchers also noted that similar AI tools with advanced capabilities are likely to emerge in the near future.
To better understand the risks, financial institutions are preparing to test the model themselves. C. S. Venkatakrishnan, chief executive of Barclays, stated that the situation requires immediate attention.
He stressed the importance of identifying exposed vulnerabilities and addressing them quickly to prevent potential disruptions.
As artificial intelligence continues to evolve, global financial leaders face increasing pressure to strengthen cybersecurity systems and adapt to new technological challenges.
