Canada’s financial watchdog has sent a clear warning to banks and other lenders: advanced AI can shorten the time available to contain cyber risk. The Office of the Superintendent of Financial Institutions, or OSFI, flagged Anthropic’s Claude Mythos as an example of how quickly that pressure could build.
The alert was delivered by email to a wide audience across the financial sector, including banks, insurers, and leaders responsible for technology, information security, and risk management. In comments to Reuters, OSFI said its concern is not the technology itself, but how institutions govern and manage the risks created by its use.
Why the regulator is concerned
OSFI said the bulletin it circulated includes guidance and practices designed to help financial institutions identify, mitigate, and respond to risks faster. The regulator said it is taking a technology-neutral approach while focusing on the risks that come with new tools, including advanced AI models such as Mythos.
| Issue | Who It Affects | What OSFI Says |
|---|---|---|
| Advanced AI models | Banks and financial institutions | The window for effective risk mitigation may become shorter |
| OSFI bulletin | Financial sector | Provides guidance to speed up identification, mitigation, and response |
| Email notice | Technology, information security, risk management, and insurance teams | Sent beyond banks alone |
Talks were already underway in April
According to Reuters, concerns about Mythos in Canada were already being discussed in early April. The meeting brought together several Canadian bank executives and regulators, showing that attention on AI use in finance had been building for some time.
The Canadian discussion also followed strong warnings from US Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell. They were said to have met bank CEOs to highlight the risks tied to the model.
Banks are still expanding AI use
At the same time, Canadian banks continue to increase investment in AI. Their use cases are also shifting from experimental projects toward more practical deployments such as chatbots, internal tools, and reduced dependence on third-party tools.
That shift shows how deeply AI is moving into banking operations, but it also explains why supervisors want stronger risk controls in place. As adoption accelerates, OSFI’s warning underscores a simple point: innovation is moving fast, and mitigation needs to move just as quickly.
