Major US Banks Crack Down on Third-Party AI Agents and Shift Liability to Users

image 1120
🔊 Listen to the summary

Major financial institutions across the United States are moving quickly to shield themselves from liability surrounding automated browsing bots, warning account holders that delegating transactions to third-party AI agents is done strictly at their own risk.

Wells Fargo recently enacted an updated access agreement cautioning customers directly about the risks of third-party automation tools. That policy update follows a move by JPMorgan Chase to ban third-party agents that log in with account credentials, alongside official advisories and revised terms rolled out by Bank of America, Citibank, and PNC.

Pushback against external autonomous software is also gaining traction outside the financial sector. Online retail giant Amazon blocked Meta’s Muse agent from checking out on its storefront, while United Airlines similarly moved to prevent personal bots from purchasing passenger flights. Standing in contrast to the strict bans, American Express has explored limited buyer safeguards, indicating that certain agentic purchasing errors, such as accidental quantity overorders, may qualify for reimbursement if merchants refuse a refund.

While banks are barring outside bots from accessing their infrastructure, lenders are aggressively pouring resources into their own first-party artificial intelligence systems. Wells Fargo noted that its internal generative AI assistant has logged over one billion user interactions, while Bank of America rolled out an updated EricaAssist platform for corporate staff and Citigroup established a dedicated internal development environment for specialized agents.

Source: Wired