Major American banks are moving swiftly to shut doors to artificial intelligence (AI) agents, joining e-commerce giant Amazon in blocking these automated tools from logging into customer accounts and moving money. Citigroup and Bank of America have become the latest banking giants to formally restrict AI agents from accessing user accounts, while JPMorgan Chase CEO Jamie Dimon signaled this week that his bank will also enforce strict limits. The banking sector’s defensive pivot follows a major backlash against autonomous tools – most notably Meta’s viral personal AI assistant, Muse – sparking fears that software could initiate uncontrolled cash movements, erode vital deposit bases and leave unresolved liability questions in its wake.
The $1.6 trillion deposit threat
Behind the restrictions lies an existential commercial threat to traditional banking models: AI agents automatically sweeping idle household cash out of low-yield deposit accounts into higher-yielding alternatives to maximise returns. Such low-cost, non-interest-bearing deposits form the foundational funding base that commercial banks rely on to issue profitable loans.According to Apollo Chief Economist Torsten Slok, approximately $1.6 trillion in non-interest-bearing deposits held collectively across JPMorgan Chase, Bank of America and Wells Fargo could be jeopardised if AI agents routinely reallocate consumer cash.Automating cash migration into competitive yields could saddle lenders with an estimated $47 billion in additional annual interest costs.Regulators have largely welcomed the banks’ restrictive guardrails, though officials note that future legislation will ultimately be required to define legal authorisation standards and clarify financial liability when autonomous software executes transactions.
Wall Street leaders demand human authorisation
Bank executives argue that allowing third-party algorithms direct transactional control creates untenable security, legal, and operational risks. Citigroup laid out its policy explicitly, confirming that it does not allow “third-party AI agents or assistants to log into customer accounts or conduct transactions directly within our digital banking platform,” emphasising that the block is necessary to preserve the security of customer accounts and private financial data.Bank of America CEO Brian Moynihan similarly indicated that the institution is restricting autonomous agents from consumer deposits, while simultaneously working on protocols to define who carries ultimate responsibility when an AI triggers a payment.“We feel very strongly that there should be a human involved in authorizing before money moves out of your account,” Moynihan said in an interview this week.Moynihan pointed out that lenders cannot afford scenarios where customers dispute transactions after the fact, claiming funds were transferred without their full awareness because an automated agent acted on their behalf.
Echoing Amazon’s crackdown on Meta’s Muse, Perplexity’s Comet
Last month, Amazon moved to block Muse from browsing and executing automated purchases on Amazon.com. The retail giant implemented the technical block after unsuccessful efforts to persuade Meta to voluntarily exclude Amazon from the tool’s autonomous shopping routines.The action follows prior enforcement by Amazon, which last year sent a cease-and-desist letter to AI search engine Perplexity demanding it halt its bots and remove the e-tailer from its Comet browser integration.
