Canadian banks blocking AI agents amid security concerns

Canada’s biggest banks, including RBC, TD, CIBC, and others, are blocking AI agents from accessing online bank accounts over security concerns.

The decision comes as major U.S. banks, like Morgan Stanley, Wells Fargo, and others, reportedly detected unauthorized activity from AI agents on customer accounts, with some agents apparently taking actions that are independent of user instructions.

Earlier this year, tech companies launched AI agent tools that they claim can take action for users and help them complete tasks. Meta’s Muse garnered the most attention, but ChatGPT maker OpenAI wasn’t far behind with the launch of its own agent tool, Dots. Users could connect these tools to bank accounts or credit cards, and even ask them to shop for products, although purchases usually require human approval.

However, alongside the Muse hype came tons of reporting about the myriad security issues with the software, including a severe flaw that could have let attackers take over the AI and do whatever they wanted on the victim’s device.

Several Canadian banks confirmed to CBC News that they would block or limit third-party agents from online banking platforms, including RBC, TD, CIBC, and Scotiabank. Meanwhile, RBC’s CEO Dave McKay said in a LinkedIn post that autonomous systems increase risks for fraud, consumer harm, and the broader economy.

But beyond banking, Canadians should be incredibly cautious about what information they give to AI tools like chatbots or agents, as well as how they use those tools, amid increasing privacy and security issues. We’ve seen AI chats leaked to the open web, security issues in AI apps, AI tools doing things they shouldn’t, and more. At this point, it seems to me that using these tools just isn’t worth the risk.

Image credit: Jon Lamont (RBC, Scotiabank), Shutterstock (CIBC)

Source: CBC News