🔍 Read the full analysis: AI Agents And Your Money: Set Up Safer Access Without Sharing Passwords on ThorstenMeyerAI.com
Get hardware and tech essentials 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 tools, saying customers are responsible for activity resulting from shared access. The warning comes as financial firms confront AI agents operating in customer accounts and companies explore agent-driven payments and trading. Limited permissions and payment controls can reduce exposure, but legal responsibility and protections remain uncertain.
Bank of America has warned customers not to share their usernames or passwords with AI tools or agents, saying they are responsible for activity resulting from shared access. The warning arrives as financial companies grapple with AI agents using customer credentials and explore ways for software to make payments or trade, leaving questions about control and liability for users.
The email, titled “Protecting your finances in the AI World,” was posted by a customer on Bluesky and reported by TechCrunch on October 27. Bank of America confirmed to TechCrunch that the message was authentic. The bank wrote that credentials “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.” It also said it was “excited about what AI used responsibly can deliver.”
Axios reported on October 22, citing unnamed sources, that major U.S. banks were finding third-party AI agents inside customer accounts, generally using customers’ own passwords. The report said banks often could not tell whether activity was authorized by the account holder. Separately, Visa told investors this week that trust was a major barrier to AI agents paying for goods, while Robinhood said customers had placed more than $250 million into accounts that AI agents trade. Those developments show financial firms are confronting both risks and demand, but they do not establish that agent trading or payments are safe.
One alternative to sharing a full login is to give an agent narrower permissions: read-only access, a requirement for approval before each payment, a virtual card with a spending cap, or a separate account with limited funds. A scoped token could restrict an agent to a particular merchant, amount and time period. These controls limit what the software can do, but they do not by themselves settle who is responsible if a transaction goes wrong.
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 Full Banking Logins Matter
A banking password can give an agent access to far more than the single task a user intended. If the agent can authenticate as the account holder, the bank may have difficulty distinguishing its actions from the customer’s or from a criminal’s. That can create two different problems: an agent could carry out an unwanted action, or fraud controls could flag the account and block legitimate access.
There is also a possible mismatch between technical safeguards and legal protection. In the United States, Regulation E excludes certain transfers initiated by someone who was given access by the consumer. The source article interprets that provision as potentially relevant to AI agents, but says no regulator or court has ruled on how it applies. In the European Union, PSD2 requires customers to take reasonable steps to protect personalised security credentials. Whether giving credentials to an AI agent would amount to gross negligence has not been tested in court, according to the source. The legal outcome in either jurisdiction should not be treated as settled.
That uncertainty makes permission design important for consumers and financial firms. A small, revocable permission is easier to limit than a master login. But, as Robinhood’s reported terms illustrate, a company can add a separate account and spending limits while still placing risk on the customer: the company said customers “assume all risk” for trades by AI agents. Users need to check both what an agent can do and who bears losses before enabling it.
Products Worth Considering
Trezor Safe 7 Crypto Hardware Wallet
Trezor Safe 5 Crypto Hardware Wallet with Color Touchscreen, Black Graphite
Ledger Flex Crypto Wallet with 2.8-Inch E Ink Touchscreen
As an affiliate, we earn on qualifying purchases.
Banks Face Agent Access
The bank’s warning reflects a broader challenge: many AI agents work by interacting with websites and services on a user’s behalf. If an agent signs in with a person’s ordinary credentials, the financial institution may see a familiar login rather than a distinct, limited authorization. The October 22 Axios report said banks were blocking agents they considered unauthorized and revising terms as they tried to determine whether account activity was customer-directed.
AI agents can also be influenced by material they encounter while completing a task, including text embedded in web pages, emails or files. The source article identifies this as prompt injection and cites OWASP’s ranking of it as a leading risk for AI applications. That does not mean every agent will be manipulated, but it gives a reason to avoid granting software broad, persistent access to money or accounts.
Financial firms are testing different approaches. Read-only connections can let an agent retrieve information without initiating transfers; approval prompts put the user in the payment path; and spending limits or dedicated balances can constrain potential losses. These are different safeguards, not guarantees. Their practical protection depends on how they are implemented, what permissions they cover and whether users can revoke them.
““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
Liability Rules Remain Untested
No court or regulator has ruled, according to the source material, on whether the cited U.S. electronic-transfer rule applies to an AI agent given a customer’s credentials, or whether sharing credentials with an agent could count as gross negligence under EU payment rules. The legal analysis described in the source is the author’s reading of the laws, not a binding decision or legal advice.
Details also remain limited about how widespread bank account access by third-party agents is, which firms are affected and what specific activity the agents have carried out. The reported Robinhood figure does not specify a time window or comparison baseline in the supplied material, so it should be read as a reported amount rather than a measure of growth. The effectiveness of particular agent controls, and the terms offered by other financial providers, may differ.
Check Permissions Before Connecting
Customers considering an AI financial tool should review the permissions it requests, the actions it can take without approval, how to revoke access and the provider’s rules on losses or disputed activity. Where available, a user can favor restricted access, transaction-by-transaction approval, amount limits or a separate account over sharing a primary banking password. These steps reduce the scope of access but cannot promise that losses will be reimbursed.
Financial companies are likely to keep adjusting account controls and terms as they encounter more agent activity. The next meaningful signals will be clearer permission standards, disclosures from providers about liability, and any regulator or court guidance on how existing payment protections apply to AI agents. Until then, customers should treat a full banking login as broader access than most individual agent tasks require.
Key Questions
Did Bank of America confirm the warning email was real?
Yes. TechCrunch reported on October 27 that a Bank of America spokesperson confirmed the email, titled “Protecting your finances in the AI World,” was authentic.
Does sharing a banking password with an AI agent remove fraud protection?
That is not settled. The source article says a U.S. rule excludes certain transfers initiated by someone given access by the customer, but no regulator or court has ruled on how the provision applies to AI agents. EU rules also require customers to protect security credentials, but the effect of sharing them with an agent has not been tested in court.
What is a safer way to let an AI agent help with money?
Options include read-only access, requiring approval for each payment, setting a virtual-card spending cap, limiting an authorization to one merchant and time period, or using a separate account with limited funds. These controls can constrain access, but do not guarantee that a provider will cover a loss.
Why can an agent using my password cause problems with a bank?
The bank may see the agent’s login as yours, or as activity by someone who stole your credentials. That can make it harder to identify whether a transaction was authorized and may lead to an unwanted payment or an account lock triggered by fraud checks.
Do spending limits mean the AI provider accepts responsibility?
No. The source reports that Robinhood used a separate account and spending limits while saying customers “assume all risk” for trades by AI agents. Users should read the provider’s terms rather than assume a technical limit determines who pays.
Source: ThorstenMeyerAI.com
