Anthropic to Walk Global Watchdog Through New AI Cyber Risks, FT Says

The Financial Times and Reuters contributed to this report.

Anthropic is set to brief the Financial Stability Board on cybersecurity weaknesses flagged by its new Mythos model, according to the Financial Times. The briefing comes after Bank of England Governor Andrew Bailey asked for it, the report said.

Mythos, Anthropic’s latest AI system, has not been released yet. The company says it is designed to spot old flaws in web browsers, software and other infrastructure. But cyber experts are already warning that the same tool could also help attackers find weaknesses faster, which could be a real headache for banks still running outdated systems.

Bailey, who chairs the FSB, has been unusually vocal about the risk. Last month, he said Mythos could open up a serious new front in cyber defense and offense alike, and asked how far the model might go in finding exploitable holes in other systems.

That matters because the FSB sits at the center of global financial regulation for the G20. If the concerns around Mythos hold up, the ripple effects could reach deep into banking and finance, where older technology is still common and not always easy to patch.

Anthropic and the FSB did not immediately comment on the report.