Some of Canada’s biggest banks say they are now banning AI agents, like Meta’s Muse, while others are taking initial steps to limit them as the broader financial industry becomes increasingly concerned about potential harms to institutions and people’s wealth.
AI agents, or agentic AI, are software that acts as a personal assistant for accomplishing given tasks. A human user might give an AI agent parameters and prompts from organizing a calendar to drafting emails and making a shopping list.
However, an Axios report on Oct. 22 said some AI agents attempted to take actions that appeared to be independent of user instructions, citing people familiar with the matter.
In some cases, bank customers have shared their login ID and password with agents, with banks typically flagging and blocking those transactions, a person in the banking industry said in that reporting. Still, such transactions cannot be deemed unauthorized if credentials were shared by the customer.
That’s given rise to growing questions about whether Canadian banks are prepared for the trend.
-
![]()
Carney announces $1.6B investment in Cohere to strengthen Canada’s AI sovereignty -
![]()
3rd party AI agents banned by several of Canada’s biggest banks as fears grow -
![]()
AI data centre concerns in Texas ahead of U.S. midterm elections -
![]()
Getting Back to Business -
![]()
How much artificial intelligence should be included in art?
Global News asked Canada’s five largest commercial banks on Oct. 22 if they planned to ban AI agents, as banks in the U.S. are doing, amid these reports. All banks that responded initially deferred to the Canadian Bankers Association, which said it cannot comment on the operational matters of its members.
TD and RBC later told Global News this week that they have moved to block all third-party AI agents.
“TD has made the decision to block all third-party personal AI agents from direct access to banking information, where authentication is required, while we ensure the appropriate safeguards, authorization mechanisms and oversight are in place,” a spokesperson for TD said Wednesday in a statement.
Get weekly money news
Royal Bank of Canada (RBC) confirmed to Global News that it sent out a company memo alerting employees on Tuesday.
“RBC has decided to block third-party personal AI agents from direct access to authenticated RBC digital properties including online banking,” a spokesperson from RBC said in a statement.
The Bank of Montreal (BMO) also told Global News on Tuesday that it has taken steps to ban AI agents from accessing some of its systems.
“To protect our clients, BMO does not permit AI agents to access our digital properties or execute transactions,” a BMO spokesperson said in a statement.
BMO also referenced its security guidelines for protecting personal information, saying it will never make an unsolicited attempt to contact someone and ask for “sensitive information,” including passwords, PINs or verification codes.
It adds: “Never share this sensitive information with any unauthorized BMO channels, including through AI agents.”
A spokesperson for CIBC also responded to a Global News request Wednesday.
“We are following developments in personal AI agents closely as these technologies evolve,” a spokesperson for CIBC said in a statement.
“We do not support personal AI agents directly opening or accessing accounts. Clients should not share their banking credentials with an AI agent and should carefully consider what personal or financial information they provide to any AI service.”
Scotiabank said Wednesday that it will “limit” AI agents.
“To protect our clients and their financial information, Scotiabank will limit third-party AI agents from accessing its banking platforms as we closely monitor evolving concerns,” a spokesperson for Scotiabank said in a statement.
As the Canadian financial world takes note of the risks posed by AI agents, the federal government told Global News it’s aware of the reports stemming from the U.S.
“Our government is monitoring this issue closely, and so are banks,” a spokesperson for Canada’s minister of AI and digital innovation, Evan Solomon, said in statement on Oct. 22.
“Canadians should be extremely cautious about sharing their banking passwords with anyone, whether that’s an AI agent, an individual or a company.”
One example of an AI agent is Muse by Meta, which can take banking and payments information and make purchases on behalf of human users.
When Global News asked Meta for a statement regarding the concerns among the banking industry, a spokesperson referred to a blog post from September.
“Muse runs on its own dedicated computer in the cloud, contained so no one else’s agent can reach it. That is where Muse lives and where the data and credentials for any service a person connects are securely stored. A separate Sentinel agent runs on that same machine, kept apart from Muse at the system level. Nothing Muse does reaches the internet unless the Sentinel approves it, and it asks the person for permission when needed,” the blog from Meta said.
Amazon told Global News on Oct. 23 that it has requested Meta remove Muse from accessing its e-commerce platform on behalf of customers.
“We think it’s fairly straightforward that third-party applications that offer to make purchases on behalf of customers from other businesses should operate openly and respect service provider decisions about whether or not to participate,” said an Amazon spokesperson in a statement.
“This helps ensure a safe, secure, and reliable customer experience, and it is how others operate including food delivery apps and the restaurants they take orders for, delivery services apps and the stores they shop from, and online travel agencies and the airlines they book tickets with for customers.
Amazon added: “Agentic third-party applications such as Muse have the same obligations, and we’ve requested that Meta remove Amazon from the experience.”
– with files from Reuters and Global’s Jacqueline Hansen
Screw your ad
Well, you know it must be crooked when the most crooked of all wont engage. If banks have said no, so if its crookeder than them, its very twisted.
The PEOPLE once again are not given a voice in this. AI is just being crammed down our throats. Time to burn META and the other billionaires out!
Ben, your ABC obsession is terminal. Seek MAID. You realize your untrustworthy banks are the Liberal Party’s biggest donors. Right? Oops
Bank staff had obligations for proper handling of customer data long before AI tools. This is about staff judgement and training more than the bugbear of AI.
Ben sounds like a hypocrite. He’s telling is we cannot trust banks (no shet!) yet he gets on his knees for Mark Carney. Spin another straw man you well-poisoning bootlicker.
@ Jack – Windows does not fix itself. Microsoft programmers fix it with an automated process that they created to install their fixes. It’s not magic.
I suggest you search ‘AI software stability’ and read some of what turns up. The priests of the AI religion are telling us their god is real. Nope.
Simple way to explain HOW AI could take over the world – when people were first able to buy computers we had to use floppy disks to start up the computers every time we shut the computers down. Now not only does microsoft WINDOWS power up and FIX ITSELF unlike in the past when we all needed computer experts to fix our computer because our children put gaming disks into the computer and the com.sys file became corrupted, Not any more. When is the last time you had to phone for an expert to fix your windows run computer? I bet a long time ago because like my computer it pretty much keeps running and fixing itself so IF windows could do that what do you think AI could do to software on computers since AI is a lot smarter then most people realize.
AI, AI
@ 5:56 Anonymous – Those are certainly … words. When GST was introduced in 1991 by Mulroney it was a revenue neutral, progressive (visible in your words) replacement for the regressive (hidden in your words) manufacturing tax.
Taxes are taxes. Inflation is inflation.They are not the same. Inflation is a very poor word to use for this in spite of your confidence. Chatbot again?
Banks can’t be trusted. Out of a list of 10 where 10 is the very worst, they are the penultimate untrustworthy entity. Followed way way behind by conservative politicians at number 10.
Meaningless rebuttal from an insecure individual with no aptitude to have an objective response, or chatbot.
“Anonymous
October 28, 2026 at 5:43 pm
@ 5:33 Anonymous – What a nonsensical word salad. Did you write it with a chatbot?”
The reality is, business and any technical argument, will always prefer regulation that is clearly defined, not too onerous, and ultimately what is the most simple. Consider it akin to lightning or any electrical current following the path of least resistance.
The reason we pay GST/HST today is because we attempted to create a transparent tax rather than the tax being hidden in the cost of goods after…you guessed it, Pierre Trudeau’s terms in office.
Today, we pay GST/HST on top of other taxes hidden in the cost of our products created by Justin Trudeau’s and Mark Carney’s terms in office/so far.
There is no better word to describe this than inflation.
Absolutely – inflation is more than just that example, but the example still holds true.
@ 5:33 Anonymous – What a nonsensical word salad. Did you write it with a chatbot?
I wouldn’t be so carried away with such simplicity as a “machine”.
The mechanics part is not the concern, it is the lack of regulation.
People could forseeably be physically, psychologically, or financially ruined by AI without regulation or consequence.
It is a direct ancestor of the government that fails to consider regulation in building structures, or electrical safety standards, shipbuilding codes and the like.
Liberals are now the eternal constant mathematical reference to short-sightedness. But there’s an unmistakable smell of sinister ideologies, pun intended.
I know there is a complete lack of critical thinking today as exhibited by the likes of Insufferable Ben and far too many others, but ignorance can be fixed in a way hatred cannot.
What’s the consequence for AI and AI companies if their product commits crimes?
Nothing. I hear Liberals are working on it, but have no timeline.
I would call it oblivious if I didn’t think it was intentionally malicious.
No, not fear, sensibility. Allowing AI access to your banking is like giving a criminal your bank card PIN.
There is no law or regulation that sets out any law to remove rational concerns that unregulated technology is a security risk.
I am not giving my voice, image or personal details and if it is collected clandestine Ly you will be sued into oblivion.
Anybody that lets a machine spend their money deserves what happens to them.