Customers who give outside AI tools access to Citibank accounts are taking responsibility for transactions and even hackers’ activity under terms changed, TechCrunch reported.[1] Citibank, the third-largest commercial bank in the United States, joins other major banks setting terms for customers handing financial tasks to automated assistants.[1]
Citi’s own AI agents carry no additional liability for customers, TechCrunch reported.[1] As outside tools become more widely available, customers choosing them must weigh the convenience of delegated financial work against responsibility for what follows, including compromised credentials.[1]
AI agents can use customers’ banking credentials to analyze spending and investments, offer advice, handle transactions and pay bills, TechCrunch reported.[1] A Citi spokesperson told TechCrunch the revised agreement clarifies the risks of sharing credentials and makes customers responsible for activity through agents they authorize.[1]
JPMorgan and Bank of America announced similar policies last month, while PNC changed its terms on Nov. 18, TechCrunch reported.[1] PayPath, which supplies AI agents for financial businesses, wrote that Wells Fargo’s updated agreement, effective November 24, 2026, treats instructions from an authorized agent as directly authorized by the customer, without further verification.[2]
Banks also see a competitive threat, TechCrunch reported: tools that help customers avoid fees or low-interest deposits threaten bank margins.[1]
Anthropic’s personal finance tool, launched earlier this month, only reads information and cannot move, manage or invest money, TechCrunch reported.[1]