🔍 Read the full analysis: What Safe AI Money Management Looks Like 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 has warned customers not to give banking credentials to AI assistants, saying customers are responsible for activity resulting from shared access. The warning comes as Visa identifies trust as a barrier to AI payments and Robinhood reports more than $250 million placed in accounts for AI-agent trading. Safer approaches limit what an agent can access and require approval for payments, but who bears losses remains uncertain.
Bank of America has warned customers not to share banking passwords with AI assistants or agents, saying customers are responsible for activity resulting from that access. The confirmed email arrives as payment and investment companies move toward agent-driven financial activity, raising questions about account security, fraud protections and who pays when an agent makes a mistake.
The email, titled “Protecting your finances in the AI World,” tells customers their 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.” A customer posted an image of it on Bluesky; TechCrunch reported on the email on October 27, and a Bank of America spokesperson confirmed to the publication that it was authentic. The bank also said it was “excited about what AI used responsibly can deliver.”
The warning is not a general ban on AI tools. It focuses on giving an outside service the same credentials a customer uses to access a bank account. An agent that signs in with those details may appear to the bank as the customer, making it harder to distinguish an authorized task from a stolen login or an action the customer did not intend.
The wider financial industry is already testing agent-related services. Visa told investors that trust is the biggest barrier to AI agents making payments, according to the source material. Robinhood said customers have placed more than $250 million into accounts that AI agents trade. The report does not provide a time period or comparison baseline for that figure, so it should be treated as a reported amount, not a measure of growth.
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
Why Agent Access Changes the Risk
The central safety issue is the scope of access. A banking password can open an account to a broad range of activity, while an agent generally needs permission for a narrower task. The source’s practical recommendation is to use limited access rather than a master credential: read-only permissions where possible, explicit approval for each payment, or a virtual card with a spending cap.
Other proposed controls include a separate account or pool of funds for agent activity, and an agent-specific token restricted to a particular merchant, amount and time window. These measures can reduce the potential damage if an agent misunderstands an instruction or is manipulated by material it reads. They do not, by themselves, settle who is legally responsible for a transaction or guarantee that a bank will reimburse a loss.
That distinction matters to customers weighing convenience against financial exposure. The source reports that Robinhood has created a separate account and spending limits, but also says the company’s terms make customers “assume all risk” for trades by AI agents. Consumers need to understand both the technical safeguards and the contractual allocation of risk before enabling automated decisions.
Products Worth Considering

DJI Mini 4K Drone Combo with 4K Camera, 3-Axis Gimbal, and Two Batteries

Samsung Galaxy Tab A11+ Kids Edition, 128GB, Royal Blue

Apple Watch Series 12 GPS 42mm Smartwatch
As an affiliate, we earn on qualifying purchases.
Banks Confront Agent Logins
The warning follows a report by Axios on October 22, citing unnamed sources, that major U.S. banks were finding third-party AI agents inside customer accounts. According to that report, agents commonly used customers’ own passwords, and banks often could not tell whether the activity was requested by the customer. The source says banks have been blocking unauthorized agents and revising their terms in response.
There is also a security concern beyond ordinary mistakes. AI agents may read web pages, emails, invoices and files while carrying out a task. Malicious or misleading instructions hidden in that material can attempt to redirect an agent, a risk known as prompt injection. The source identifies this as a leading risk in the OWASP list for AI applications. It does not establish that any such attack caused the banking activity described here.
Rules on payment disputes do not offer a settled answer for every agent scenario. In the United States, Regulation E’s definition of an unauthorized electronic transfer excludes some transfers initiated by a person given account access by the consumer. The source author says applying that provision to an AI agent has not been resolved by a regulator or court. In the European Union, PSD2 requires customers to take reasonable steps to keep personalised security credentials safe; how sharing credentials with an AI agent would be assessed under those rules has not been tested in court, according to the source.
“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 its customer email
Who Bears an Agent’s Losses
No court or regulator ruling identified in the source settles how existing consumer protections apply when an AI agent uses credentials its customer supplied. The source author’s interpretation of U.S. and EU rules is not legal advice, and its account of how those rules might apply should not be treated as a definitive legal finding.
It is also unclear how consistently banks can identify whether an agent’s action was authorized, how different providers will define customer approval, or what protections will apply when a transaction is disputed. Limits and separate accounts may reduce exposure, but they do not answer those questions. The source gives no details on the duration or comparison basis for Robinhood’s reported $250 million in agent-trading accounts.
Check Permissions Before Connecting
For customers, the immediate step is to review an agent’s requested permissions and the financial service’s terms before connecting an account. Avoid sharing a main banking password; prefer limited or read-only access, set transaction caps, and require human approval for payments or trades where the service allows it. Customers should also check how to revoke access and report activity they did not authorize.
For companies building agent payments, the source’s safety checklist points toward narrowly scoped credentials, clear approval steps, spending limits and a separate source of funds. Providers will also need to explain who is responsible when an agent acts unexpectedly. Banks’ evolving policies and any future regulatory or court decisions may clarify the liability questions, but no specific ruling or industry timetable is identified in the source material.
Key Questions
What did Bank of America tell customers about AI agents?
It warned customers not to share their banking username or password with outside services, including AI tools and agents. The email said customers are responsible for activity resulting from shared access.
Is it safe to give an AI agent a banking password?
The bank’s guidance is not to share the password. A login can grant broad access and may make an agent’s activity look like the customer’s, complicating security checks and disputes.
What are safer ways to let an agent handle money?
Where available, use read-only permissions, approval for each payment, a capped virtual card, a separate account, or a token limited by merchant, amount and time. These controls can limit exposure but do not guarantee reimbursement.
Will a bank reimburse a loss caused by an AI agent?
That remains uncertain and may depend on the facts, the provider’s terms and applicable law. The source says no regulator or court has settled how the cited U.S. and EU rules apply specifically to AI agents using customer-provided credentials.
Source: ThorstenMeyerAI.com
