Major U.S. banks have detected unauthorized AI agents accessing customers’ accounts, raising concerns about these systems’ ability to carry out financial transactions beyond users’ instructions, according to a report by Axios cited by Al Arabiya Business.
The report said that banks including Morgan Stanley, Bank of America, JPMorgan Chase, and Wells Fargo are racing to contain the risks associated with these agents after cases of unauthorized access to their systems were detected.
The problem often arises when customers use AI-powered agents or assistants to access their accounts using their personal login credentials. Although customers may consent to this, most banks require account holders to log in themselves, without relying on third parties or automated systems.
Security concerns are mounting as these agents may be granted permission to make purchases or transfer money, creating operational and legal risks that are difficult to contain. Some banks have also observed AI agents contacting customer service centers or attempting to carry out actions that do not clearly align with users’ original instructions.
Banking officials have described the situation as “extremely concerning,” amid fears that these systems could access customers’ accounts and take actions affecting their financial circumstances without direct human supervision.
These risks have prompted several banks to tighten restrictions on external agents’ access to their systems. Morgan Stanley has begun blocking third-party AI agents from accessing its platforms, while other institutions, including JPMorgan Chase, Bank of America, and Citigroup, have taken similar steps.
Meanwhile, a spokesperson for Chase emphasized that customers should not share their usernames and passwords with any other parties. A Morgan Stanley spokesperson, however, told Axios that the bank had not identified any cases of “rogue” AI agents gaining access to its platforms.
Most banks have moved toward a compromise that allows AI tools to access customers’ financial information through “read-only” permissions, without enabling them to conduct transactions or change account details. This approach resembles the mechanisms used by external financial applications, such as budgeting and expense-tracking software, which can view data without controlling accounts.
Industry officials have acknowledged that demand for this technology is expected to grow in the coming years, putting banks under pressure to balance innovation with security protection.
Financial institutions are struggling to distinguish between activities carried out directly by customers and those performed by AI agents, as well as to verify whether these agents have proper authorization or are accessing accounts through illegitimate means.
Despite substantial investments in protection systems and cybersecurity, officials believe the rapid advancement of AI technologies is exposing new vulnerabilities at a pace that exceeds banks’ ability to address them. Institutions are also preparing for emergency scenarios through plans designed to secure customers’ accounts or restore access quickly in the event of large-scale incidents involving AI agents.
United News Network – UNN Arabic
An independent media platform providing reliable news and objective analysis, seeking to promote peace and cultural dialogue around the world, convey the truth, and build bridges of understanding among peoples.
For more news, visit our homepage:
Author: Counselor Faisal Al-Mutairi.
Publication Date: October 25, 2026.
Updated: October 2026.
