Banks are beginning to consider how much access artificial intelligence (AI) agents should have to customers’ online banking accounts, as financial institutions weigh the potential of agentic AI against security and account-access concerns.
In a recently released survey of 70 bankers, the Federal Reserve Bank of Dallas found that 11.4% said their institution or service provider had taken steps to restrict existing or prospective customers’ AI agent activity on online banking platforms. Another 22.9% said they planned to introduce restrictions within the next six months.
The findings suggest that banks are still developing policies for a technology that could change how customers interact with financial services. More than a quarter of respondents, 27.1%, said they did not know whether their institution or service provider had restricted AI-agent activity. Meanwhile, 38.6% said no restrictions had been implemented and were not planned.
The results highlight an emerging challenge for banks as AI agents gain the ability to navigate digital platforms and perform tasks on behalf of consumers. Financial institutions are now in the position of balancing customer convenience and emerging AI capabilities with concerns about authentication, authorization, fraud and the security of account information. Banks may need to distinguish between agents acting with explicit customer permission and automated activity that could create new vulnerabilities.
The findings come as financial institutions face broader cybersecurity and fraud concerns. In the same survey, 44.3% of bankers identified cybersecurity as one of their institution’s three leading concerns over the next six months, while 41.4% cited fraud.
The Dallas Fed collected responses between Oct. 27 and Nov. 4 from bankers in its Eleventh Federal Reserve District. The survey noted that the results provide a regional snapshot of how financial institutions are approaching AI-agent access rather than a measure of policies across the U.S. banking industry.
In October, PYMNTS reported that banks are cracking down on unauthorized AI agents. Major U.S. banks are detecting and deleting unauthorized third-party AI agents that are acting within customer accounts, with some agents attempting to take actions that were not requested by the user.
According to a separate PYMNTS report also published in October, JPMorganChase, Citigroup, Bank of America and Charles Schwab all said that AI agents will not be allowed to move customer money.
