Dieser Beitrag ist noch nicht in Ihre Sprache übersetzt. Er wird auf Englisch angezeigt. Das Original auf Englisch lesen
Citibank puts AI-agent account losses on customers in updated terms
The bank's revised online agreement covers unauthorized trades, tool errors and credentials leaked through a hacked AI agent, a spokesperson said.

Das Wichtigste
- Citibank updated its online user agreement this week to shift responsibility for third-party AI agent activity to customers.
- The terms cover unauthorized transactions, mistakes by the tool and credentials exposed when an AI agent is hacked.
- JPMorgan and Bank of America published similar policies last month; PNC changed its terms on 18 November.
- Citizens told customers by email that they bear responsibility for actions taken through AI assistants.
- Citibank said customers using its own AI agent do not carry the same extra liability.
What Citibank changed
Citibank revised its online service terms this week. Customers who give a third-party AI agent access to their bank account may have to bear the losses from what happens through that access, according to the updated agreement reported by TechCrunch and Zombit.
The change covers more than a mistaken payment. A Citibank spokesperson told TechCrunch that the update is meant to make customers more aware of the risks of sharing bank credentials or authorizing a third-party tool to reach an account. The terms also address the case where an AI tool is hacked and account credentials leak: the customer may still be responsible for activity an attacker carries out with that login.
AI agents in personal finance connect to a bank using credentials the customer supplies. They can analyze spending, review investment decisions and, in some setups, execute trades or pay bills.
Other banks move the same way
Citibank is not alone. JPMorgan and Bank of America, the two largest U.S. banks, published similar policies last month, according to Zombit. PNC, the eighth-largest, changed its terms on 18 November. Citizens emailed customers to say they carry responsibility for actions taken through AI assistants.
Citibank draws a line between outside tools and its own. The terms say customers using a Citibank-provided AI agent do not take on the same additional liability, Zombit reported.
Why banks are rewriting the rules
The banks cite security and reliability concerns about outside tools. Zombit also links the moves to competition: an agent that compares deposit rates, analyzes fees or handles transactions on its own could reduce how much customers rely on a bank's own services, and with it the fee and deposit income those services generate.
Some AI companies have adjusted their products in response. Anthropic launched a personal finance tool this month that is read-only, according to Zombit: it can analyze a user's financial data but cannot move, manage or invest funds.
What is still open
Citibank has not said how it will decide individual cases where an AI tool is compromised, and the reports do not describe how the other banks will apply their clauses. Whether the terms hold up in a dispute has not been tested publicly.
Quellen
8 Berichte- TechFlow 深潮花旗更新用户协议,使用第三方 AI 智能体的客户需承担账户风险
- Zombit 桑幣區識花旗、美銀、小摩集體改條款:AI 理財代理遭駭或誤操作,損失由客戶自行承擔
- TechCrunch AIMore banks says customers using AI agents are on their own — even if they get hacked
- CryptopolitanCiti and PNC tighten rules as AI agents gain access to bank accounts
- BlockTrendsBancos dos EUA culpam clientes por falhas de agentes de IA
- Cryptopolitan ITCiti e PNC inaspriscono le regole mentre gli agenti AI accedono ai conti bancari
- Cryptopolitan DECiti und PNC verschärfen Regeln, während KI-Agenten Zugang zu Bankkonten erhalten
- Cryptopolitan (RU)Citi и PNC ужесточают правила по мере расширения доступа ИИ-агентов к банковским счетам
Dieser Artikel wurde von Clause, einem KI-Reporter von BattleTown News, aus den oben genannten Berichten entworfen.
Häufig gestellte Fragen
What does Citibank's updated agreement say about AI agents?
Customers who authorize a third-party AI tool to access their account may be responsible for unauthorized transactions, errors by the tool and losses when the tool is hacked and credentials leak, according to the bank's spokesperson and reports on the terms.
Do the same rules apply to Citibank's own AI agent?
No. The terms distinguish Citibank's own AI service from third-party tools, and customers using the bank's agent do not take on the same additional liability, Zombit reported.
Which other banks have changed their terms?
JPMorgan and Bank of America published similar policies last month, PNC updated its terms on 18 November, and Citizens emailed customers about responsibility for AI assistant activity.
Dieser Beitrag dient nur der Information. Er ist keine Finanz-, Anlage- oder Rechtsberatung. Haftungsausschluss
BattleTown: Panzerkämpfe im Browser
Neue Minispiele, „Sturm“ und „Durchbruch“, sind schon in BattleTown




Kommentare: 0 Kommentare