Canada’s banking regulator, the Office of the Superintendent of Financial Institutions, warned the country’s biggest banks and insurers about cyber risks tied to Anthropic’s Claude Mythos and other advanced AI models, according to Reuters. OSFI told chief technology, security, and risk officers in an April email that such models sharply shorten the window to catch and patch vulnerabilities, per records Reuters obtained via a freedom-of-information request. Concern reaches past Canada, with US Treasury Secretary Scott Bessent and then-Fed chair Jerome Powell holding an emergency call with bank CEOs about the model’s cyber threat. OSFI put out public guidance on generative and agentic AI on Monday after Reuters raised questions, stressing that it cares how firms manage AI risk rather than about the tools themselves.
