🔍 Read the full analysis: How To Keep AI Money Tasks Separate From Your Bank Password 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 their banking username or password with AI tools, saying customers are responsible for activity resulting from shared access. The warning comes as financial firms confront AI agents using customer credentials and some companies offer agent-based trading or payments. Limited access and transaction approvals can reduce exposure, but the legal and financial responsibility for an agent’s actions remains unsettled.
Bank of America warned customers not to give their banking username and password to an AI assistant, saying they are responsible for activity resulting from shared access. The warning comes as banks deal with third-party agents accessing customer accounts and financial companies test AI-driven payments and trading.
The email, titled “Protecting your finances in the AI World,” says credentials should not be shared with another person or an outside service, including an AI-enabled tool or agent acting on a customer’s behalf. A customer posted the email on Bluesky, TechCrunch reported on October 27, and a Bank of America spokesperson confirmed to TechCrunch that it was authentic.
The bank’s message says that customers who share credentials are responsible for activity resulting from that access. It also says Bank of America is “excited about what AI used responsibly can deliver.” The email does not spell out how the bank would handle specific losses tied to an AI agent or define what counts as responsible use.
Other developments show that financial activity involving AI agents is already taking place. Visa told investors this week that trust is a major barrier to agents making payments, while Robinhood said customers had put more than $250 million into accounts that AI agents trade. That amount is a reported account total; the source material does not provide a comparison period or baseline.
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 Logins Give Agents Too Much Access
A banking password can let software act with the same access a customer has, rather than granting permission for one narrowly defined task. That can make it difficult for a bank to distinguish a customer-directed action from a stolen login or an agent’s mistake. The potential consequences include an unwanted transfer, a blocked account or a dispute over who bears the loss.
Sharing credentials may also affect protections, but the outcome depends on the facts and applicable rules. In the United States, Regulation E excludes certain transfers made by someone the customer furnished with account access, unless the customer has notified the bank that the person is no longer authorized. Applying that provision to an AI agent or the company operating it is an interpretation, not a settled legal ruling.
In the European Union, PSD2 requires customers to take reasonable steps to keep personalized security credentials safe. The source material says whether giving those credentials to an AI agent would count as gross negligence has not been tested in court. A bank’s warning is not itself a legal decision about liability, so customers should not assume that ordinary fraud protections will resolve every agent-related dispute.
Security controls can narrow an agent’s ability to act, but they do not automatically settle responsibility. Robinhood, for example, has described separate accounts and spending limits for agent trading, while also saying customers “assume all risk” for those trades, according to the source material. That makes the terms governing an agent service as relevant as its technical limits.
Products Worth Considering
Dell XPS 13 9350 Laptop, 13.4 FHD+ Display, Intel Ultra 7 256V, 16GB DDR5, 1TB SSD
Logitech MX Master 4 for Mac, Ergonomic Bluetooth Mouse with Haptic Feedback, Ultra-Fast Scrolling, USB-C Charging, for MacOS, iPadOS, Windo
Dell XPS 13 9350 AI Professional Laptop 13.4 120Hz FHD+ Display
As an affiliate, we earn on qualifying purchases.
Banks Are Seeing Agents Use Customer Credentials
According to an Axios report on October 22, citing unnamed sources, major U.S. banks were finding third-party AI agents in customer accounts, often logged in with the customer’s own password. The report said banks may have trouble determining whether the customer intended the activity. This account is attributed to Axios and its sources; the material provided here does not include a public bank-by-bank accounting.
When an agent uses a customer’s login, the bank’s systems may see activity that appears to come from the customer or from someone with stolen credentials. Banks may respond by blocking access or changing terms, but those actions can also affect customers who authorized an agent. The source material says financial firms are working to distinguish user-directed agent activity from unauthorized access.
One alternative to sharing a full login is to give an agent a narrower form of permission: read-only access, approval required for each payment, a virtual card with a spending cap, or a token restricted to a merchant, amount and time period. A separate account funded with only a limited amount can also cap potential losses. These are possible safeguards, not guarantees against errors, fraud or disputes.
Liability Rules Have Not Been Tested
No court or regulator ruling cited in the source material determines whether giving an AI agent banking credentials falls within the relevant U.S. electronic-transfer rules or amounts to gross negligence under EU payment rules. Those legal applications remain interpretations, not established outcomes. This article is not legal advice.
It is also unclear how banks will treat individual cases, including whether a customer explicitly approved a particular transaction, whether an agent exceeded its instructions, or whether the agent’s provider contributed to the loss. The source material does not detail the scope of Robinhood’s account limits or provide a full set of terms for every agent-payment service.
Reports of banks finding third-party agents are based on unnamed sources, and the material does not identify the institutions, number of affected accounts or total losses. The reported $250 million at Robinhood is not accompanied by a stated time window or comparison baseline. Those limits prevent a fuller measure of how widespread agent-based money activity is.
Check Permissions Before Connecting an Agent
Customers considering an AI money tool can review what it can access before connecting an account. Prefer permissions limited to the task, such as read-only access for budgeting or a requirement to approve each transaction. If a payment feature offers a virtual card, spending cap, merchant restriction or time limit, check that those restrictions are active and understand how they can be changed.
Read the provider’s terms for responsibility when an agent makes an unwanted transaction, and check the bank’s account rules before granting access. Avoid entering a bank password directly into an agent or service unless the bank and service clearly describe a secure, limited authorization method. A separate, low-balance account can limit exposure, though it cannot prevent every loss or resolve liability questions.
For companies building agent payment systems, the source material points to narrower credentials, transaction-level approval and clear limits as alternatives to handing agents a customer’s full login. The next developments to watch are how banks change their account terms and controls, how payment providers define agent permissions, and whether regulators or courts clarify responsibility when an authorized AI agent causes a loss.
Key Questions
Should I give an AI agent my bank password?
Bank of America says not to share it with an outside service or AI-enabled agent. A full login can give software broad access and may make it harder to distinguish its activity from yours.
What is a safer way to let an AI handle a money task?
Use permissions limited to the task where available, such as read-only access, a requirement to approve each payment, or a virtual card with a spending cap. Check the exact scope and duration of access before enabling it.
Will my bank reimburse an AI agent’s unwanted transaction?
That is not settled by the material cited here. Rules may depend on the type of transfer, the customer’s actions and the applicable law. Bank of America’s email says customers are responsible for activity resulting from shared credentials, but it does not decide every possible claim.
Can an AI agent be manipulated into making a payment?
The source material says agents can be steered by instructions hidden in webpages, emails or files, a risk known as prompt injection. This is one reason to restrict an agent’s permissions and require approval for transactions.
Do spending limits remove the risk of AI trading?
No. Limits can cap some activity, but they do not guarantee that an agent will act as intended or determine who pays for a loss. Robinhood’s stated terms, as reported in the source material, say customers assume all risk for trades by AI agents.
Source: ThorstenMeyerAI.com
