🔍 Read the full analysis: Thinking Of Letting AI Handle Money? Protect Your Credentials on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- 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 banking passwords with AI tools or agents, saying customers are responsible for activity resulting from that access. Visa identified trust as a barrier to agent-led payments, while Robinhood said customers had placed more than $250 million in accounts used for AI-agent trading. The legal and financial responsibility for an agent’s actions remains uncertain.
Bank of America has warned customers not to share their usernames or passwords with AI-enabled tools or agents, saying customers are responsible for activity resulting from that access. The warning comes as Visa cites trust as a barrier to agent-led payments and Robinhood says customers have put more than $250 million into accounts where AI agents trade.
The bank’s email, titled “Protecting your finances in the AI World,” says credentials should never be shared with another person or provided to an outside service, including an AI tool or agent acting on a customer’s behalf. A customer posted a screenshot of the email on Bluesky; TechCrunch reported it on October 27, and a Bank of America spokesperson confirmed its authenticity to the outlet.
The bank also says it is “excited about what AI used responsibly can deliver.” Its warning is specifically about giving an outside service direct access through a customer’s credentials. Once an agent logs in with a person’s password, the bank may have difficulty distinguishing the agent’s actions from the customer’s or from activity by someone who stole the password.
The source report says Visa told investors this week that trust is the biggest barrier to AI agents making payments. Robinhood said customers had placed more than $250 million into accounts that AI agents trade. That figure describes funds in those accounts; the source does not say how many customers are involved, the measurement period, or how much the agents have traded.
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 Bank Login Access Raises Stakes
A banking password can grant access well beyond a single requested task. If an agent is given the same login a customer uses, the access may allow it to view account information and initiate actions as that customer, rather than operate under a narrow permission for one payment. That makes credential sharing a question of both security and responsibility, not only convenience.
The potential consequences include an unwanted transaction, a compromised account, or a bank blocking access after treating an agent’s activity as suspicious. The source report cites US and EU rules that place duties on account holders, but it also acknowledges that applying those rules to AI agents has not been tested by a regulator or court. Customers should not assume that ordinary fraud protections will necessarily apply when they have deliberately given software their login.
The activity described by Visa and Robinhood shows why this issue is pressing: financial services are already being used in connection with agent payments and trading. More agent access does not by itself establish that the systems are unsafe, but it increases the importance of clear permission limits, records of approval and an explanation of who bears losses when something goes wrong.
Popular right now
Ninja Foodi 10-Quart DualZone 6-in-1 Air Fryer, Black, DZ302
De'Longhi Rivelia Automatic Espresso Machine with Milk Frother – White
Vitamix Propel Series 750 Professional-Grade Blender, 4 Programs, Black
As an affiliate, we earn on qualifying purchases.
Banks Respond to Agent Logins
On October 22, Axios reported, citing unnamed sources, that major US banks were finding third-party AI agents inside customer accounts, generally using the customer’s own password. The report said banks often could not tell whether the activity was authorized by the customer. It also said banks were blocking some agents and revising their terms. Those details come from Axios’s reporting and unnamed sources, rather than a public statement from every bank involved.
The source article describes two relevant legal frameworks. In the United States, Regulation E excludes from its definition of an unauthorized electronic fund transfer one initiated by a person given access by the consumer, subject to the rule’s conditions. In the European Union, the Payment Services Directive requires users to take reasonable steps to keep personalized security credentials safe. How either framework applies when credentials are shared with an AI agent is unresolved; the article’s author characterizes that application as a personal reading, not a court or regulator ruling.
The source recommends limiting what an agent can do rather than handing it a full account login. Options it lists include read-only access, approval for each payment, a capped virtual card, or a narrowly scoped token restricted by merchant, amount and time. A separate account or pot of money can also limit exposure, though limits do not settle who is liable for losses.
““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 as reported by TechCrunch
Liability Rules Remain Untested
No court or regulator ruling identified in the source settles how existing consumer protections apply when a customer gives banking credentials to an AI agent or to the company operating it. The source’s discussion of US and EU rules is an interpretation, not legal advice or an official determination. The effect may depend on the facts, applicable terms and jurisdiction.
It is also unclear from the source how many banks have detected agent logins, how widespread those cases are, or what specific transactions were involved. Robinhood’s reported $250 million figure lacks a stated measurement window and comparison baseline, and it does not establish the amount of trading or any gains or losses. The source does not provide the terms or technical details for Visa’s agent-payment plans.
Technical safeguards also cannot guarantee that an agent will act as intended. The article notes that instructions hidden in web pages, emails or files can attempt to influence an agent, a form of attack commonly called prompt injection. It does not identify a particular incident involving a customer’s bank account, so the risk should not be confused with evidence that a specific attack has occurred.
Use Narrow Permissions Before Connecting
For customers considering an AI agent, the source’s practical guidance is to avoid sharing a primary banking password and first check whether the provider and bank offer a limited access method. Prefer permissions that restrict an agent to a specific task, require approval before money moves, or set a hard amount and time limit. Read the service’s terms to understand who is responsible for transactions before enabling automatic approvals.
For banks and companies developing agent payments, the open questions include how to distinguish customer-approved activity from unauthorized access, how to stop or reverse a mistaken payment, and how to explain responsibility in plain language. Visa’s comments indicate trust remains a challenge, while Robinhood’s use of a separate account and spending limits offers one example of controls described in the source. Those controls are not a substitute for clear liability terms.
Further clarity may come through updated bank policies, product disclosures or regulatory and court decisions. Until then, customers should treat a full account password as broad authority, not a convenient shortcut—and verify the permissions and risk terms of any service before connecting it to money.
Key Questions
What did Bank of America tell customers?
It warned customers not to give their usernames or passwords to outside services, including AI tools and agents. The email said customers are responsible for activity resulting from shared credentials.
Does sharing a password mean a customer always loses fraud protection?
That is not settled by the source. It discusses US and EU rules that may be relevant but says their application to AI agents has not been tested by a court or regulator. The outcome may depend on the circumstances and applicable terms.
What is a safer way to let an AI agent handle a payment?
The source suggests limited permissions, such as read-only access, approval for each payment, a capped virtual card, or a token limited to a merchant, amount and time window. A separate account can also limit the funds exposed.
What does Robinhood’s $250 million figure measure?
The source says customers had put more than $250 million into accounts that AI agents trade. It does not give the measurement period, number of customers, amount traded, or a comparison baseline.
Source: ThorstenMeyerAI.com
