🔍 Read the full analysis: The Simple Guide To AI Agents And Safer Money Handling on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Bank of America warned customers not to give banking credentials to AI tools, saying customers are responsible for activity resulting from shared access. The warning comes as banks grapple with third-party agents entering accounts and financial companies test agent-driven payments and trading. Safer controls can limit access, but who bears losses from an agent’s actions remains unsettled.
Bank of America warned customers this week not to give their usernames or passwords to AI assistants, saying customers are responsible for activity resulting from that access. The email, confirmed as authentic by a bank spokesperson to TechCrunch, comes as financial companies confront AI agents using customer credentials and test systems that let software handle payments or trades.
The email, titled “Protecting your finances in the AI World,” says banking credentials should never be shared with another person or an outside service, including an AI-enabled tool acting on a customer’s behalf. The bank also said it is “excited about what AI used responsibly can deliver.” A customer posted the email on Bluesky, and TechCrunch reported on it on October 27; the publication said a Bank of America spokesperson confirmed it was genuine.
The warning lands amid wider activity around agent-led finance. According to the source report, Visa told investors trust is a major barrier to AI agents making payments, while Robinhood said customers had put more than $250 million into accounts that AI agents trade. Axios reported on October 22, citing unnamed sources, that major U.S. banks were finding third-party AI agents inside customer accounts, often logged in with customers’ passwords. Banks may not be able to distinguish customer-directed agent activity from access by someone else, the report said.
One alternative to handing over a full login is to give an agent limited permissions: read-only access, payment-by-payment approval, a capped virtual card, or a token restricted to a particular merchant, amount and period. A separate account with a limited balance can also contain potential losses. These measures reduce what an agent can do, but they do not by themselves settle who is legally responsible if it makes a mistake or acts on harmful instructions.
Before your agent touches money
Give it a valet key, not the master key
Logging in with your password makes the agent you. A credential made for the agent can do only the job you gave it, and you can take it back.
- What it opensEverything you can do: transfers, payees, settings, statements
- LimitsNone that the agent cannot get around by logging in as you
- What the bank seesYou, or a thief. It cannot tell which
- If a web page tricks itThe trick reaches your whole account
- To take it backChange your password everywhere it was used
- If money goes missingUS rules can exclude transfers by someone you gave access to; Bank of America: “you’re responsible”
- What it opensOnly what you allowed: read-only, one card, one pot of money
- LimitsBuilt in: amount, merchant, time window, approval on your phone
- What the bank seesAn identified agent acting for an identified person
- If a web page tricks itThe damage stops at the limit
- To take it backDisconnect the agent or delete its card. Your password never moved
- If money goes missingDepends on the terms. Read them first: some still put all risk on you
A comparison of two set-ups, not a measurement. Which limits exist depends on your bank, card and app.
Sources: Bank of America customer email, confirmed to TechCrunch (27 Oct 2026); 12 CFR 1005.2(m); OWASP LLM01. Comparison: Thorsten Meyer.
Powered by Thorsten Meyer AI
Who Bears the Cost of Agent Errors
The central issue is not simply whether an AI agent can complete a transaction. It is what authority the customer gives it and who bears the loss if the agent pays the wrong recipient, follows deceptive instructions or makes an unintended trade. When an agent uses a customer’s password, the bank may see an ordinary customer login rather than a distinct software actor. That can make it harder to investigate disputed activity and may lead to account restrictions.
Rules that protect consumers from unauthorized transfers may not clearly fit activity carried out through credentials the customer voluntarily shared. The source author interprets U.S. and EU rules as creating potential risks for customers who give an agent account access, but says the application to AI agents has not been decided by courts or regulators. Readers should treat that legal discussion as an interpretation, not a definitive ruling or individual legal advice.
Robinhood’s reported use of a separate account and spending limits illustrates the difference between technical safeguards and liability. The source says Robinhood tells customers they “assume all risk” for trades by AI agents. A cap can limit exposure, but customers still need to understand the provider’s terms before allowing an agent to act without individual approval.
Products Worth Considering
DJI Osmo Pocket 3 Filter Bundle – Vlogging Camera with 1 CMOS Sensor and 4K/120fps
Amazon Kindle Scribe Colorsoft 64GB (1st Generation)
Apple iPad Air (2025) 13‑inch Wi‑Fi 256 GB – Space Gray (M3 chip)
As an affiliate, we earn on qualifying purchases.
The concern has developed as AI assistants move from answering questions to taking actions across websites and services. To carry out a financial task, an agent may encounter webpages, emails, invoices or files that contain instructions. The source describes prompt injection as text in those materials that can steer an agent away from its intended task, and says OWASP ranks it as a leading risk for AI applications. The source does not provide a specific incident involving a bank agent being redirected this way.
Giving an agent a full banking password is especially broad because it can make the software appear to the bank as the customer. A more restricted permission can narrow the task and potentially make activity easier to contain. The practical distinction is between sharing a master credential and granting a limited capability, such as viewing balances or paying a named merchant up to a stated amount.
The source also cites consumer-protection rules, while cautioning that their application to AI agents is untested. In the United States, it discusses Regulation E’s treatment of transfers initiated by someone given access by a consumer. In the European Union, it points to PSD2 obligations to safeguard personalised security credentials. Whether those provisions apply to an AI agent or its operator in a particular case is not established in the material.
““Your username and password should never be shared with another person or provided to an outside service, including an AI-enabled tool or agent that offers to act on your behalf.””
— Bank of America, in the customer email reported by TechCrunch
Unsettled Rules and Agent Controls
No court or regulator ruling is identified in the source that determines whether giving an AI agent banking credentials counts as furnishing access under U.S. transfer rules, or whether doing so breaches EU credential-safety duties. The outcome may depend on the facts, the account terms and the law that applies.
It is also unclear how widely banks are encountering third-party agents, which services or types of account access are involved, and how banks distinguish a customer-authorized action from a compromised login. The Axios account cited unnamed sources, and the source material does not give figures for affected customers or transactions.
Specific safeguards vary by provider. The source offers examples of restricted access but does not establish that every bank or AI service supports them, or that any one control prevents all losses. Customers should check what permissions an agent receives, whether a human must approve payments and what the provider’s terms say about disputed activity.
Check Permissions Before Connecting
Customers considering an AI finance tool can start by not sharing their primary banking password. They can ask whether the service offers read-only access, a transaction-specific approval step, a limited virtual card or a capped separate account. They should also read the terms governing unauthorized or agent-initiated activity and contact their bank about permitted connection methods.
For financial firms and agent developers, the developments point to a need for clearer limits on what an agent can do, visible approval records and a way to revoke access. The source gives no timetable for new industry standards, regulatory guidance or changes to bank policies. The next concrete developments will be whether providers expand restricted-access options and clarify how responsibility is handled when an agent acts.
Key Questions
What did Bank of America tell customers about AI agents?
The bank said customers should not share usernames or passwords with outside services, including AI tools or agents. It also said customers are responsible for activity resulting from shared credentials.
Does sharing a password definitely remove fraud protection?
That is not established by the source. It describes possible legal complications under U.S. and EU rules but says their application to AI agents has not been tested in court. The bank’s email states its position; it does not settle every legal or disputed transaction.
What is a safer way to let an AI agent handle money?
Use the narrowest available permission, such as read-only access, approval for each payment, a capped virtual card, a restricted token or a separate account with limited funds. Availability depends on the bank and service.
What is prompt injection?
It is an attempt to steer an AI system through instructions placed in material it reads, such as a webpage, email or file. The source identifies it as a risk for agents that process outside content, but does not report a specific bank incident caused by it.
What should customers check before connecting an agent?
Check the permissions requested, whether payments require approval, how access can be revoked and what the service’s terms say about losses from agent activity. Avoid giving an agent a primary banking password.
Source: ThorstenMeyerAI.com
