Canadian Banks Block Third-Party AI Agents
AI-generated image · Bay Street Wire
RBC, TD, BMO, and CIBC restrict tools like Meta's Muse from accessing client accounts, citing security risks.
Major Canadian financial institutions are blocking or limiting the ability of third-party personal AI agents to directly access client online banking accounts and sensitive information, according to CBC News.
Royal Bank of Canada (RBC), TD Bank, CIBC, and Bank of Montreal (BMO) have all stated that these restrictions are intended to protect client data. RBC spokesperson Jeff Lanthier told CBC News that the bank's decision to block these applications from its online platforms followed a careful weighing of risks. TD spokesperson Ada Lovelace Rampersaud added that the bank will block agents where authentication is required, emphasizing that clients should not share passwords or PINs with AI tools. CIBC spokesperson Deborah Rowe stated the bank does not support agents directly opening or accessing accounts.
While these banks have invested billions in AI for internal operations, such as mortgage processing, their leadership has expressed skepticism regarding external tools. RBC CEO Dave McKay wrote in a LinkedIn post that as systems become more autonomous, risks including fraud and consumer harm increase. BMO's CEO has also warned against providing banking information to AI agents.
These tools, including OpenAI's ChatGPT and Meta's Muse, typically allow authorized users to monitor transaction histories and balances. However, Axios reported that major U.S. banks—including JPMorgan, Bank of America, Morgan Stanley, and Wells Fargo—have detected unauthorized AI agents operating within customer accounts, some of which took actions independent of user instructions. Scotiabank has also indicated it will limit these agents while monitoring the situation.