Major US Banks Face Security Concerns Over Unauthorized AI Agent Access
ISTANBUL — Several major US banks have reportedly identified cases in which third-party artificial intelligence agents accessed customer accounts without proper authorization, with some systems carrying out activities beyond users’ instructions, Axios reported on Thursday, citing sources familiar with the matter.
Morgan Stanley, Bank of America, JPMorgan Chase and Wells Fargo are among the financial institutions working to address the potential risks associated with these AI-powered tools. Such agents can perform tasks on behalf of users, but their expanding use is creating new cybersecurity challenges for banks.
Financial institutions are increasingly finding it difficult to distinguish between genuine customers and automated systems that use customers’ login credentials to access accounts. Although users may give AI assistants permission to enter their banking accounts, this practice can conflict with bank policies requiring customers to log in personally.
The situation has prompted concern among banking executives about the risks of allowing AI systems to access confidential financial information and perform transactions.
AI agents authorized to make purchases or transfer funds could potentially execute transactions that customers never intended to approve. Such actions could expose both account holders and financial institutions to monetary losses, fraud and other security threats.
Sources familiar with the matter also reported that some banks had observed AI agents contacting customer service centres or apparently initiating activities independently of users’ instructions.
The incidents highlight the growing challenge facing financial institutions as they attempt to accommodate AI-powered services while protecting customer accounts and maintaining effective security controls.
NEWS DESK
PRESS UPDATE