BMO blocks use of third-party AI agents on its sites

Bank of Montreal BMO-T has blocked third-party artificial intelligence agents from accessing its digital properties and executing transactions, the latest move in a string of banks that have restricted the technology.

BMO made the decision to bar the technology last week. Royal Bank of Canada RY-T, the country’s biggest bank, said Tuesday it would block third-party personal AI agents from directly accessing the bank’s digital properties, following similar moves by U.S. lenders.

The move is aimed at safeguarding clients, BMO spokesperson Jeff Roman said in a statement.

“BMO uses AI responsibly to deliver value for our clients while protecting their information and data,” Mr. Roman said.

“Clients should never share sensitive banking information, including passwords, PINs or verification codes, through unauthorized channels, including AI agents.”

BMO chief executive Darryl White said at an event last week that customers should not share sensitive banking data with AI agents. Using AI to compile a grocery or to-do list or to do a job more efficiently is acceptable, but the agents will not protect sensitive information including banking usernames and passwords, he said.

“Somebody said to me, ‘It’s kind of like if you put a sticky note inside the windshield of your car in a big parking lot at a mall and walked away for a few hours. Would you be comfortable putting all that stuff there where everybody walks by and reads it?’” Mr. White said. “You’re not going to do that. Don’t do it here.”

Customers can authorize AI agents – semi-autonomous AI entities designed to carry out instructions from humans such as Meta Platforms Inc.’s META-Q popular AI agent Muse – to access their bank accounts, transaction history and data to analyze spending patterns and make recommendations.

National Bank of Canada NA-T said Wednesday it is closely monitoring the development of consumer-facing AI agents, including Muse, and assessing the opportunities and associated risks the technology may pose to clients.

“Our first duty is to preserve trust and protect our clients’ personal information,” National Bank spokesperson Alexandre Guay said in a statement. “We welcome innovation that delivers value to clients while preserving security and privacy.”

Lenders are investing heavily in AI to boost productivity and generate revenue. But the surging adoption of AI has also prompted concerns over cybersecurity and financial crime, causing banks and regulators to bolster oversight of the technology.

U.S.-based banks JPMorgan Chase & Co., Bank of America Corp., Morgan Stanley and Charles Schwab Corp. have previously blocked or limited outside AI agents’ access to their customers’ data.

The Financial Stability Board – a global industry watchdog – has recommended that financial institutions reinforce oversight of their AI systems, including banning agents’ access to sensitive information.

Canada’s top banking regulator said Tuesday he is ramping up oversight of AI practices at the country’s banks, but would not restrict them from allowing customers to use AI agents.

Superintendent of Financial Institutions Peter Routledge said he would leave the decision on AI agents up to individual banks, provided they adhere to the regulator’s guidelines and maintain financial stability. The Office of the Superintendent of Financial Institutions has been developing tighter safety guidelines on AI systems.

Mr. Routledge said customers could choose to switch financial services providers based on whether or not they want to use these types of services.

“I don’t think we would get down to the level of you should do it this way, and so everyone should do it this way,” Mr. Routledge said.