🔍 Read the full analysis: Use AI To Help Manage Money While Keeping Bank Access Limited on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Bank of America warned customers not to share their banking usernames and passwords with AI tools, saying customers are responsible for activity resulting from that access. The warning comes as payment companies and trading platforms report growing use of AI agents, while the legal and financial consequences of agent mistakes remain unsettled.
Bank of America warned customers this week not to share their banking usernames or passwords with AI tools, saying they are responsible for activity resulting from access they provide. The warning comes as Visa cited trust as a barrier to AI agents making payments and Robinhood said customers had placed more than $250 million into accounts AI agents trade, raising questions about how to limit access and who pays when an agent makes a mistake.
The 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-enabled tool or agent acting on a customer’s behalf. A customer posted the email on Bluesky, TechCrunch reported it on October 27, and a Bank of America spokesperson confirmed its authenticity to the publication. The bank also said it is “excited about what AI used responsibly can deliver.”
The bank’s warning is about the scope of access as well as the possibility of fraud. An AI agent that signs in using a customer’s ordinary password may be able to do what that customer can do, and the bank may have difficulty distinguishing the agent’s activity from the customer’s or from someone who stole the credentials. That could lead to unwanted transfers or trading, or to a bank blocking access while it investigates activity it considers suspicious.
There are more limited ways to give software financial access, though their availability depends on the provider. These include read-only permissions, requiring customer approval for each payment, setting limits on a virtual card, or using a separate account with a restricted balance. A scoped payment token could be limited to a particular merchant, amount and time period. These controls can reduce what an agent can do; they do not by themselves settle who bears legal responsibility if something 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
The Cost of Giving Agents Access
The practical risk is that a person may hand over more authority than a task requires. A password is a general-purpose credential, not permission for one specific purchase. If the agent misreads an instruction, acts on manipulated information, or is steered by text it encounters online, broad account access could expose money well beyond the intended transaction.
The legal position is not settled. In the United States, Regulation E defines an unauthorized electronic transfer in a way that excludes certain transfers initiated by a person given access by the consumer. Whether that provision applies to an AI agent or the company operating it is not established by the source material; the author notes that no regulator or court had ruled on that question as of publication. In the European Union, PSD2 requires customers to take reasonable steps to protect their personalised security credentials, and sets out possible liability where that duty is breached intentionally or through gross negligence. Whether sharing a password with an AI agent meets that standard is also untested.
That uncertainty matters to customers deciding whether to enable an agent’s financial features. Technical limits and legal responsibility are separate issues: a service may add spending caps or a separate account while still placing some risk on its users. Customers should read the applicable terms before granting access, particularly before turning off approval steps.
Popular right now

Ninja Luxe Café Premier 3-in-1 Espresso Machine, Drip Coffee & Cold Brew

Vitamix 5200 Blender with 64 oz Container, Black

KitchenAid Artisan Mini Tilt-Head Stand Mixer, 3.5 Quart, Aqua Sky Blue
As an affiliate, we earn on qualifying purchases.
Banks Face Unclear Agent Activity
The warning follows a broader concern about third-party AI tools accessing financial accounts with customers’ own login details. Axios reported on October 22, citing unnamed sources, that major U.S. banks were finding unauthorized third-party agents inside customer accounts. The report said banks often could not tell whether the activity reflected what the customer wanted. The account in the supplied source does not identify the banks or give figures for the reported activity.
Separately, Visa told investors that trust is the biggest barrier to AI agents that pay for goods and services. Robinhood said customers had put more than $250 million into accounts traded by AI agents. The source does not specify the measurement period, the number of customers, or a comparison baseline for that figure, so it should be read as a reported total rather than a rate of growth.
AI agents can also encounter instructions inside webpages, emails or documents that attempt to redirect their behavior, a risk commonly called prompt injection. The source identifies it as a leading risk in the OWASP list for AI applications. Limiting an agent’s financial permissions can reduce the possible impact of an error or manipulation, but does not establish that the agent will always follow the user’s intent.
“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
It is not clear how consumer-protection rules would apply if an AI agent uses credentials supplied by an account holder and initiates a transfer the holder did not intend. The source’s legal discussion is the author’s interpretation, not a court ruling or regulator guidance. The outcome could depend on the facts, the relevant jurisdiction, the account agreement and the way access was granted.
The source also does not provide the full terms of the financial products mentioned, explain how Robinhood’s account safeguards work in practice, or state the time period behind the $250 million figure. Bank of America’s email does not specify which protections might apply to a particular customer after credentials are shared. No detailed incident figures or confirmed losses tied to the reported bank activity are given.
Check Permissions Before Connecting
Customers considering an AI financial tool can check whether it offers read-only access, transaction-by-transaction approval, spending caps or a separate, limited-balance account. They can also review the provider’s terms for who is responsible for unauthorized activity and how to revoke access. A password should not be treated as a narrow permission for one task.
Financial institutions and payment providers may continue to adjust how they handle agent activity as use expands. The source reports that banks are blocking some unauthorized agents and revising terms, but it does not name institutions, provide dates for policy changes, or set out a common industry standard. For now, the permissions available and the liability terms vary by service, and the legal treatment of AI-agent access remains unresolved.
Key Questions
What did Bank of America tell customers?
It warned customers not to share their usernames or passwords with outside services, including AI-enabled tools and agents. It said customers are responsible for activity resulting from access they provide.
Why is giving an AI agent a bank password risky?
A standard password can grant broad account access, rather than permission for one task. The bank may also have difficulty distinguishing an agent using the customer’s login from the customer or a credential thief.
Are customers legally responsible for every AI-agent transaction?
That is not established. The source describes unsettled legal questions in both the United States and the EU. Applicable law, account terms and the details of the access may matter.
How can customers limit an agent’s access?
Where available, they can use read-only permissions, require approval for each payment, set a limited virtual-card amount, use a scoped token or keep the agent’s funds in a separate account with a restricted balance.
Source: ThorstenMeyerAI.com
