🔍 Read the full analysis: Bank Passwords And AI Agents: A Guide To Safer Money Handling 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 warned customers not to share banking usernames or passwords with AI tools, saying customers are responsible for activity resulting from that access. The warning comes as companies explore AI-driven payments and trading, while legal protections and the risks of agent access remain unsettled.
Bank of America warned customers this week not to give their banking usernames and passwords to AI assistants, saying customers are responsible for activity resulting from access they provide. The warning comes as Visa discusses trust concerns around AI agents making payments and Robinhood says customers have placed more than $250 million in accounts that AI agents trade, underscoring the importance of limiting what agents can reach.
The email, titled “Protecting your finances in the AI World,” was posted by a customer on Bluesky. TechCrunch reported on it October 27, and a Bank of America spokesperson confirmed to the outlet that the email was authentic. The bank said 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 added that customers who share them are responsible for activity resulting from that access.
Bank of America also said it is “excited about what AI used responsibly can deliver.” Its warning addresses the difference between using an AI tool for assistance and giving software the credentials needed to act inside a bank account, a distinction tied to safer AI delegation. An agent logging in with a customer’s password may be difficult for the bank to distinguish from the customer or from someone who stole the password.
The source report says Visa told investors that trust is a major barrier to AI agents paying for goods and services. Robinhood said customers had put more than $250 million into accounts traded by AI agents. Those figures indicate that agent-assisted financial activity is already being discussed and used, but the source provides no time window or comparison baseline for Robinhood’s figure.
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 Risks of Agent Bank Access
Giving an agent a bank password can give it the same broad account access the customer has, rather than permission for a single task. That may expose customers to unwanted transfers, purchases or account changes. It may also prompt a bank to block activity it cannot verify, potentially locking the customer out while the activity is investigated.
Responsibility and reimbursement are unsettled. The source author interprets US and EU rules as raising risks when a customer deliberately gives credentials to software. But that interpretation has not been tested in a court or, according to the source, resolved by a regulator. Customers should not treat the warning as a definitive ruling that every agent-related loss will be their responsibility; the outcome could depend on the facts, applicable law and bank terms.
For financial institutions and companies building agent-payment products, the issue is whether an agent can be granted enough authority to complete a task without receiving unrestricted access. Limits on what an agent can view or spend could reduce the potential harm if the system makes a mistake or is manipulated. Those technical safeguards, however, do not by themselves settle who bears the loss.
Products Worth Considering
Trezor Safe 5 Crypto Hardware Wallet with Color Touchscreen, Black Graphite
Trezor Safe 7 Crypto Hardware Wallet
Ledger Flex + Folio Crypto Hardware Wallet
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. The agents were reportedly logging in with customers’ own passwords. The report said banks were having difficulty distinguishing activity the customer intended from activity that was not authorized.
That account is separate from Bank of America’s confirmed email: the bank verified the email’s authenticity to TechCrunch, while the claims about banks encountering agents inside accounts were attributed to Axios’s unnamed sources. The source material also describes prompt injection, in which instructions hidden in a web page, email or file can influence an AI agent. It identifies this as a leading risk in OWASP’s list of AI application risks.
Safer access designs described in the source include read-only permissions, approval for each payment, spending-limited virtual cards, and tokens limited by merchant, amount and time. A separate account or pool of funds may also cap potential losses. These approaches narrow an agent’s authority compared with sharing a main banking password, though the source does not establish that any one design eliminates risk.
““If you do share your credentials, keep in mind that you’re responsible for activity resulting from that access.””
— Bank of America, in the customer email
Liability Rules Remain Untested
No court or regulator ruling is identified in the source material that determines how consumer-protection rules apply when an AI agent uses credentials supplied by an account holder. The source author’s reading of US Regulation E and EU payment rules is presented as an interpretation, not settled law or legal advice. Whether a particular customer is reimbursed may depend on the circumstances and the applicable account terms and law.
It is also unclear how many customers are using agents with bank credentials, how often such access leads to disputed transactions, and what specific controls banks can reliably use to identify agent activity. The source’s account of banks finding agents relies on unnamed sources cited by Axios. Robinhood’s reported $250 million figure has no stated measurement period or comparison baseline in the material.
Permission Controls and Bank Policies
Customers considering financial AI tools will need to check what access they request, what actions they can take, and what happens if a transaction is disputed. The source’s suggested safeguards favor narrow permissions and human approval over sharing a main account password. Customers should also review the service’s terms and their bank’s policies before connecting an account.
Banks and payment companies are likely to keep addressing how they identify and limit agent activity as these products develop. The source material does not name a timetable for new Bank of America rules, regulatory guidance or court decisions. For now, the confirmed development is the bank’s warning; the legal treatment of losses tied to customer-authorized agents remains unresolved.
Key Questions
What did Bank of America tell customers?
The bank warned customers not to share usernames or passwords with outside services, including AI-enabled tools or agents. It said customers who share credentials are responsible for activity resulting from that access.
Does the warning mean a customer always loses money taken by an AI agent?
No such universal rule is established by the source. The source author says relevant US and EU legal questions have not been tested in court and presents their reading as an interpretation, not a legal ruling.
What are safer ways to give an AI agent financial access?
The source describes read-only permissions, approval for each payment, spending-limited virtual cards, and tokens restricted by merchant, amount and time. These can limit access compared with sharing a full banking password, but they do not guarantee that an agent will be error-free.
What is prompt injection?
Prompt injection is an attempt to steer an AI agent through instructions placed in material it reads, such as a web page, email or file. The source identifies it as a major AI application risk.
What is still unknown about AI agents and bank accounts?
The legal responsibility for losses involving customer-authorized agents has not been settled in the source material. It also gives no reliable count of affected customers or disputed transactions, and does not specify the time period behind Robinhood’s reported $250 million figure.
Source: ThorstenMeyerAI.com
