Major US banks limit protection for customers who use AI tools to manage money

Exercise 1
Vocabulary
Repeat each word, its meaning and the example sentence out loud.
terms of service
noun phrase/tɜːmz əv ˈsɜːrvɪs/the written agreement that explains the rules and conditions customers must follow when using a company's product or service
Before connecting your password to an AI tool, always read the bank's updated terms of service to understand your liability.
credentials
noun/krəˈdenʃəlz/login information like usernames and passwords that give access to accounts or systems
The company policy now forbids sharing your bank credentials with any third-party AI application without explicit approval.
liable
adjective/ˈlaɪəbəl/legally responsible for paying money if something goes wrong
The bank said they are no longer liable for fraudulent transactions made by AI agents on customer accounts.
margin
noun/ˈmɑːrdʒɪn/the profit that a company makes on a product or service after paying costs
Banks worry that AI tools could cut their margins by helping customers switch to accounts with higher interest rates.
leak
verb/liːk/to accidentally allow secret or sensitive information to escape or become known to others
If the AI platform leaks your banking credentials, the bank says the customer is responsible for any unauthorized transactions.
Exercise 2
Article
Read the article aloud on your own, or listen to it first.
Major US banks limit protection for customers who use AI tools to manage money
Tap any word to see what it means.
Major US banks have changed their customer agreements to shift responsibility away from themselves when third-party AI tools access bank accounts.
Citibank, the third-largest commercial bank in the United States, updated its terms this week to tell customers that if they grant AI agents access to their accounts, any problems that result are entirely their responsibility. This includes losses from hacked credentials, failed transactions, or incorrect financial decisions made by the AI — even if the customer never explicitly authorized those actions.
JPMorgan Chase and Bank of America, the country's two largest banks, announced the same policy last month. PNC Bank, the eighth-largest, made similar changes today.
Citizens Bank also emailed customers this week with identical warnings.
AI agents for personal finance can be powerful tools. They analyze spending patterns, make investment suggestions, and can even execute transactions and pay bills if given the right access.
Some AI tools are designed to do this work automatically.
However, banks view these systems with suspicion. They worry about unreliability, security gaps, and the risk that hackers could exploit accounts where credentials have been shared with third-party tools.
A bank spokesperson explained that they wanted customers to "understand the risks associated with sharing banking credentials" and needed clarity on liability.
The banks' stance goes further than simple warnings. Citi's new terms explicitly state that if an AI tool itself is hacked or leaks credentials, the customer remains responsible for any resulting theft or fraud.
This puts the burden entirely on individual users rather than on the AI companies or the banks themselves.
Anthropic, one major AI company, took a notably different approach. The personal finance tool it launched this month is explicitly read-only — it can analyze accounts but cannot move money, make trades, or change settings.
This design eliminates many of the risks banks fear.
Beyond security, banks also view AI finance tools as competitive threats. These systems can help customers spot unnecessary fees and avoid low-interest savings accounts, which cut into bank profits.
If AI agents become a standard way customers manage money, banks risk losing control over key customer interactions.
This pattern extends beyond banking. Many service providers across industries are now pushing back against third-party AI access to their systems, creating friction that could shape which AI investments succeed and which ones face serious obstacles.
Discuss this story out loud with our AI speaking coach.
It asks one question at a time, at your level, and gives you feedback at the end.
Start the discussion Retell it3,000+ learners
Also works inChatGPTandClaude
Exercise 3
Discussion
Have a discussion based on the following questions.
- 1.Have you or your team considered using an AI tool to manage business or personal finances? What would concern you most based on these new bank policies?
- 2.If you were building an AI finance tool, how would you address the liability issues and security concerns that banks are raising?
- 3.Do you think banks are right to refuse liability when AI tools access their systems, or should they share responsibility for security?
- 4.What approach similar to Anthropic's read-only design could reduce friction between AI companies and other service providers in different industries?
- 5.How might these bank policies affect the growth of AI startups in the finance sector over the next two years?
Exercise 4
Further discussion
Have a discussion based on the following questions.
- 1.Beyond finance, what other industries do you think will push back hardest against third-party AI access, and why?
- 2.Should there be industry standards or government regulation about when companies can limit AI access to their systems?
- 3.As AI agents become more capable and reliable, should liability rules shift back toward companies and banks, or stay focused on users?
- 4.What business models might emerge for AI companies in finance if they cannot process transactions directly but only provide analysis and advice?
Discuss this story out loud with our AI speaking coach.
It asks one question at a time, at your level, and gives you feedback at the end.
Start the discussion Retell it3,000+ learners
Also works inChatGPTandClaude
Source
This lesson is our own retelling, for English learners, of a story from TechCrunch.
Related stories

Signal stops users taking screenshots to protect privacy from AI
22 November 2026 · B1

Signal Blocks AI Assistants from Reading Your Messages
21 November 2026 · B2

Major US Systems Go Down: Cyber Attack or Bad Luck?
21 November 2026 · B1

Researcher Exposes Massive Data Leaks Across AI Supply Chain Using Digital Tripwires
21 November 2026 · C1



