🔍 Read the full analysis: Can Your AI Agent Handle Money Without Access To Your Bank Login? on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Bank of America has warned customers not to share their banking usernames or passwords with AI tools, saying customers are responsible for activity resulting from that access. The warning comes as banks report concerns about third-party agents entering customer accounts, while Visa and Robinhood describe growing activity involving AI and payments. Safer options may limit an agent’s access, but their availability and legal protections vary.
Bank of America has warned customers not to share their banking usernames and passwords with AI tools or agents, saying they are responsible for activity resulting from that access. The email, confirmed as authentic by a bank spokesperson to TechCrunch, arrives as financial firms address agents entering accounts and handling money in ways that can be difficult to distinguish from actions by the customer.
The email, titled “Protecting your finances in the AI World,” said credentials should never be shared with another person or given to an outside service, including an AI-enabled tool or agent acting on a customer’s behalf. It added that customers who share their credentials are responsible for activity resulting from that access. The bank also said it is “excited about what AI used responsibly can deliver.”
TechCrunch reported on the email on October 27 after a customer posted it on Bluesky; a Bank of America spokesperson confirmed it was authentic. The warning does not establish how the bank would handle every dispute involving an AI agent. Rather, it sets out the bank’s position on customers sharing credentials and the responsibility that may follow.
On the same Tuesday as the reported warning, Visa told investors that trust is the biggest barrier to AI agents paying for goods and services, according to the source material. Robinhood said customers had put more than $250 million into accounts that AI agents trade. That figure describes the amount placed in those accounts; the material provides no comparison period or baseline for interpreting it as a rate 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 Payment Risk
A banking password is not a limited permission for one payment. It can let an agent access account information and carry out actions under the customer’s identity. That creates a basic problem: the bank may see the customer’s login, not a separate, clearly identifiable agent. The account holder could face an unwanted transaction, or a bank could flag the activity as suspicious and restrict access.
There is also a potential gap between technical safeguards and legal responsibility. The source article interprets US and EU rules as creating possible risks for customers who provide an agent with account access, but it explicitly says no regulator or court has yet ruled on how those rules apply to AI agents. The specific outcome would depend on the facts and applicable law. Customers should not treat a spending limit or separate account as a guarantee that a bank, broker or agent provider will absorb losses.
As agents take on financial tasks, questions include what an agent can access, which transactions require approval and who is responsible if something goes wrong. Controls and clear accountability are among the issues involved in the adoption of agents for payments.
Products Worth Considering
Bose QuietComfort Wireless Bluetooth Headphones with Active Noise Cancellation, Mic, USB‑C Charging, Deep Bass, 24‑Hour Battery – White Smok
Jackery HomePower 3000 Portable Power Station with Two 200W Solar Panels
Bose QuietComfort Wireless Noise-Cancelling Over-Ear Headphones, Black
As an affiliate, we earn on qualifying purchases.
Banks Face Unclear Agent Logins
The warning follows an October 22 Axios report, cited in the source material, that major US banks were finding third-party AI agents inside customer accounts, often logged in with the customer’s own password. The report said banks were concerned they could not reliably tell whether the activity was authorized by the customer. It also reported that banks were blocking some agents and revising terms.
That account is attributed to Axios and unnamed sources; it is not a public finding that every bank has experienced the same activity. The source material also points to prompt injection as a separate risk: instructions hidden in a web page, email or file can potentially influence an agent reading that content. The OWASP project lists prompt injection among prominent risks for AI applications, but the material does not establish that it caused a particular banking loss.
Possible alternatives to sharing a full login include read-only access, requiring approval for each payment, a capped virtual card, a permission limited to one merchant or amount, or a separate account with a restricted balance. These are presented as safer design patterns, not as protections available from every bank or payment service. Limited access reduces exposure, but it does not by itself settle who bears a loss.
“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
Liability Rules Remain Untested
It is not yet clear how banks will resolve disputes over payments initiated by AI agents using customer-provided credentials. The source article says that, in the United States, Regulation E excludes certain transfers initiated by someone furnished access by the consumer, but applying that provision to an AI agent or its operator is the author’s interpretation; no regulator or court ruling is cited.
In the European Union, the source points to PSD2 duties to protect personalized security credentials and possible liability in cases involving intent or gross negligence. Whether giving credentials to an AI agent would meet that standard has not been tested in court, according to the article. These are not settled legal conclusions, and the source cautions that its discussion is not legal advice.
Other open questions include which permissions specific agent products actually receive, how well approval and spending controls work, and what recourse a customer has when an agent makes an unintended transaction. The reported Robinhood account total also has no stated time window or comparison baseline in the source material.
Watch for Permission and Liability Terms
Customers considering an AI service that handles money can check its permissions and terms before connecting an account. Relevant questions include whether access is read-only or can move money, whether each payment requires approval, whether limits can be set, and what the provider says about losses. A separate, capped account may restrict the amount at risk, but it does not guarantee reimbursement.
Financial firms face questions about how to distinguish customer-directed agent activity from stolen credentials and how disputed transactions are handled. Visa’s comments point to trust as a barrier, while Bank of America’s email warns customers about responsibility when credentials are shared. Further bank policies, product controls and legal rulings may clarify how these arrangements work; the source material does not give a timetable for them.
Key Questions
Yes. In an email confirmed as authentic by a bank spokesperson to TechCrunch, Bank of America said customers should not provide credentials to outside services, including AI-enabled tools or agents.
Does the warning mean customers always pay for an AI agent’s mistakes?
No universal outcome is established in the source material. The bank said customers are responsible for activity resulting from shared credentials, but how a particular dispute is resolved may depend on the facts, account terms and applicable law.
What is a safer way to let an AI agent handle a payment?
Options described in the source include read-only access, approval for each payment, a virtual card with a spending cap, narrowly scoped payment permissions or a separate account with limited funds. Availability and protections vary by provider.
Can an AI agent be steered by information it reads online?
Potentially. Prompt injection refers to instructions placed in content such as web pages, emails or files that may influence an AI agent. The source identifies this as a risk, not as evidence that a specific banking incident occurred this way.
Source: ThorstenMeyerAI.com
