🔍 Read the full analysis: How To Delegate Financial Tasks To AI With Less Risk on ThorstenMeyerAI.com
Get smart everyday buys 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 has warned customers not to share banking credentials with AI tools, saying customers are responsible for activity resulting from that access. The report recommends limiting agents to specific tasks and funds, but legal responsibility and protections for agent-driven transactions remain unsettled.
Bank of America has warned customers not to give their banking usernames or passwords to AI agents, saying customers who share credentials are responsible for activity resulting from that access. The warning comes as financial companies report increasing contact between AI agents and money, raising questions about how to limit an agent’s authority and who bears the loss if it acts incorrectly.
The bank’s email, titled “Protecting your finances in the AI World,” says credentials should not be shared with another person or provided to an outside service, including an AI-enabled tool or agent acting on a customer’s behalf. A customer posted the email on Bluesky, and TechCrunch reported on it October 27. A Bank of America spokesperson confirmed to TechCrunch that the message was authentic.
The email says that if customers share their credentials, “you’re responsible for activity resulting from that access.” It also says the bank is “excited about what AI used responsibly can deliver.” The report argues that the warning matters because an agent logging in with a customer’s password may be difficult for a bank to distinguish from the customer or from an intruder.
Other developments point to agents already reaching financial accounts. The source report says Visa told investors that trust is the biggest barrier to AI agents making payments. It also reports that Robinhood said customers had put more than $250 million into accounts that AI agents trade. Robinhood’s safeguards, according to the report, include a separate account and spending limits; the company also says customers “assume all risk” for trades made by AI agents.
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
Limit an Agent’s Financial Reach
The practical issue is that a password usually grants broad account access, not just permission to complete one task. An agent with a customer’s login may be able to do more than the customer intended, while the bank may see activity under the customer’s own credentials. That can complicate both fraud detection and disputes over whether a transaction was authorized.
The report’s safer approach is to give an agent the smallest permission needed: read-only access when it only needs to review information; human approval before payments; or a virtual card with a spending cap. Where available, a restricted token could be limited to a particular merchant, amount and time period. Another option is a separate account or “pot” containing only money the customer is prepared to expose to the agent.
These controls can limit potential damage, but they do not settle who is liable. The source specifically warns that technical safeguards are not the same as legal protection. Customers should read the terms for an agent and its financial provider, including who bears responsibility if it makes an unwanted trade or payment.
Products Worth Considering
Logitech MX Mechanical Full-Size Wireless Mechanical Keyboard, Tactile Quiet Switches, Backlit Keys, Bluetooth, USB-C, Multi-Device, Graphit
Logitech MX Master 4 for Mac, Ergonomic Bluetooth Mouse with Haptic Feedback, Ultra-Fast Scrolling, USB-C Charging, for MacOS, iPadOS, Windo
HP Sprocket Photo Booth Machine Instant Color Photo Printer with 10.1 Touchscreen, LED Ring Light, 250+ Frames, 3×4 ZINK Paper, QR Code Shar
As an affiliate, we earn on qualifying purchases.
Banks Face Unclear Agent Activity
The warning follows a report by Axios on October 22, which cited unnamed sources who said major U.S. banks were finding third-party AI agents inside customer accounts. According to that reporting, agents generally accessed the accounts using customers’ own passwords, and banks often could not determine whether the activity reflected the customer’s instructions. The source article says banks have been blocking unauthorized agents and revising terms in response.
The article also points to prompt injection as a separate risk: an agent may read a webpage, email or document containing instructions designed to redirect its behavior. OWASP lists prompt injection as a leading risk for AI applications, according to the source. An agent that can initiate financial actions could be exposed to those instructions while processing content for an apparently ordinary task.
Legal protections vary by jurisdiction, and their application to AI agents has not been settled in court, according to the source author. The article interprets U.S. Regulation E’s exclusion for transfers initiated by someone given access by a consumer as potentially relevant to agents, while acknowledging that no regulator or court has ruled on that application. It also cites the EU Payment Services Directive’s duty to protect security credentials, but says whether sharing a password with an agent would amount to gross negligence has not been tested in court.
““If you do share your credentials, keep in mind that you’re responsible for activity resulting from that access.””
— Bank of America, in its customer email
Liability Rules Remain Untested
It is not yet clear how existing consumer protections apply when an AI agent uses credentials its customer supplied. The source author’s reading of U.S. and EU rules is an interpretation, not a court ruling or regulator’s guidance. Whether a customer’s action counts as authorization, negligence or a transfer by an entrusted party may depend on the facts and the applicable terms.
The source material does not establish how many customers are using agents to manage money, how often agent actions have caused losses, or whether banks will offer standardized restricted access. It also does not specify the terms governing every agent, account or payment product. The reported Robinhood balance is a figure attributed to the company, not a measure of losses or of the broader market.
Watch for Safer Access Options
Customers considering an AI tool for financial tasks can check whether it supports read-only permissions, transaction-by-transaction approval, spending caps, restricted cards or tokens, and a separate funded account. They should review the agent’s terms and the financial institution’s rules before connecting an account, and avoid giving an agent a primary banking password when a narrower permission is available.
For companies, the next steps are likely to include clearer permission settings and disclosures about responsibility for agent actions, but the source does not identify a timetable or a specific industry standard. Banks’ responses to account activity, updates to customer terms, and any regulator or court decisions will help clarify what protections apply. Until then, the scope of access and the allocation of risk remain unsettled.
Key Questions
What did Bank of America tell customers about AI agents?
It warned customers not to share banking usernames or passwords with people or outside services, including AI-enabled tools and agents. The email says customers are responsible for activity resulting from credentials they share.
Why is sharing a banking password with an agent risky?
A password can give an agent broad access, rather than permission for one specific task. The bank may also have difficulty distinguishing an agent using the customer’s login from the customer or someone using stolen credentials.
What are safer ways to let an AI agent handle a financial task?
Use the narrowest available access: read-only permissions, approval before each payment, a capped virtual card, a restricted token, or a separate account with limited funds. These controls can reduce exposure but do not by themselves determine liability.
Are customers legally responsible for every transaction an AI agent makes?
That is not settled by the source material. The article discusses possible relevance of U.S. and EU rules but says their application to AI agents has not been tested by a court or resolved by a regulator.
Source: ThorstenMeyerAI.com
