Bank of Montreal CEO Darryl White has cautioned consumers against providing sensitive financial information to artificial intelligence assistants, raising concerns about privacy and security as AI technology develops faster than regulatory safeguards.
Speaking at the Toronto Global Forum on Wednesday, October 21, White warned that growing familiarity with online transactions has made some consumers increasingly comfortable sharing personal information with digital platforms, potentially exposing them to security risks.
He specifically advised individuals against entering banking usernames, passwords, credit card details, social insurance numbers or personal identification numbers into AI agents, emphasizing that such information may not be adequately protected.
White also stressed the need for greater public awareness about the risks associated with emerging AI technologies, particularly as more consumers begin using automated assistants for financial tasks.
His comments followed remarks from Royal Bank of Canada CEO Dave McKay, who revealed Tuesday that RBC is considering restricting Meta’s AI assistant, Muse, from accessing customer banking information.
McKay said the bank is examining the issue carefully, seeking to balance consumer convenience with the need to protect financial data and maintain the security of its systems. He indicated that RBC would avoid making a hasty decision.
AI assistants, including Meta’s Muse and OpenAI’s ChatGPT, can connect to certain financial accounts with user authorization, allowing them to review balances, examine transaction histories and provide spending recommendations. However, financial transactions generally require additional approval from users.
The debate comes as Canada’s major financial institutions continue investing billions of dollars in artificial intelligence to improve operational efficiency and customer services.
BMO has also expanded its own use of AI. In August, White highlighted how the bank’s insurance division was using the technology to complete certain underwriting decisions in approximately 10 seconds, compared with a previous industry benchmark of 28 days.
Despite these advancements, banking executives are increasingly questioning whether third-party AI platforms should be permitted to access sensitive customer information.
Similar concerns have emerged in the United States, where the chief executives of JPMorgan Chase and Charles Schwab recently announced restrictions on external AI agents accessing client data.
The warnings highlight the growing challenge facing financial institutions as they seek to embrace artificial intelligence while ensuring that customer privacy, cybersecurity and regulatory protections keep pace with technological developments.
