RBC and BMO Chiefs Raise Alarm Over AI Agent Access to Customer Accounts
RBC and BMO Chiefs Raise Alarm Over AI Agent Access to Customer Accounts #
The chief executives of Royal Bank of Canada and Bank of Montreal, Canada’s two largest banks by market capitalisation, have each issued public warnings about the risks of allowing autonomous software to access customer financial data, with their remarks coming within 24 hours of each other.
RBC CEO Dave McKay told Bloomberg Television on Tuesday that the bank is weighing whether to block Meta’s Muse AI agent from accessing client information held at the lender. “It’s something I’m thinking about, absolutely,” McKay said, noting that no final decision had been reached. He said protecting customers carries direct liability for the bank.
The following day at the Toronto Global Forum, BMO CEO Darryl White warned consumers directly never to hand banking usernames, passwords, PINs, credit card details, or social insurance numbers to AI agents, saying such credentials are not adequately protected once shared. “I think we’ve got a ways to go on educating people,” he said, pointing to two decades of e-commerce that has made consumers broadly comfortable sharing personal data online.
RBC’s head of artificial intelligence, Bruce Ross, told The Logic separately that the firm would not permit agents like Muse to connect to its systems without sufficient safeguards. Ross also raised the unresolved question of who bears legal responsibility when an AI agent acts on a customer’s behalf and something goes wrong, a gap regulators have not yet addressed.
Both executives’ concerns are shared by peers at other institutions. JPMorgan CEO Jamie Dimon has already barred third-party agents from accessing customer accounts, warning of risks from “rogue agents.” Charles Schwab is taking a more permissive but controlled approach, planning to route third-party AI tools through Model Context Protocol servers that grant defined, limited permissions while blocking higher-risk actions such as fund transfers.
Regulators have also been paying attention. The Financial Stability Board, with backing from Canada’s Office of the Superintendent of Financial Institutions, has urged lenders to tighten oversight of AI systems, including verifying agents’ identities and restricting their exposure to sensitive customer information.
Canada’s major banks have continued investing in their own internal AI programmes. BMO has deployed AI in its insurance division to compress underwriting decisions from weeks to seconds, while RBC expects to generate up to one billion Canadian dollars in enterprise value from AI by fiscal 2027.
Copy link