Canadian financial regulators warn banks about cyber risks and have referenced Anthropic’s “Claude Mythos” in internal communications, according to reports drawing on an email. Multiple outlets say the regulator’s message links emerging AI-related threats to broader cybersecurity concerns for banks, urging institutions to consider potential vulnerabilities that could arise from how such systems are accessed, used, or integrated into banking environments. The cited materials indicate the regulator is specifically alerting banks to the possibility that advanced AI tools could be misused or could introduce new attack pathways, including risks associated with data handling and operational security. The reports do not provide evidence that the AI system itself has been compromised; instead, they describe the warning as a risk-management prompt that uses “Claude Mythos” as an example of a type of technology regulators are scrutinizing. The communications are presented as part of the regulator’s broader supervisory approach to cyber resilience across the financial sector. As described in the coverage, banks are expected to assess relevant controls, monitoring, and incident preparedness in response to the regulator’s guidance.