🔍 Read the full analysis: Set Up Your AI Agent To Handle Money Without Sharing Your Password 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 their banking usernames or passwords with AI tools, saying customers are responsible for activity resulting from shared access. The warning comes as financial firms report growing use of AI agents for payments and trading, while safer, limited-access options and customer liability remain unsettled.
Bank of America has warned customers not to give their banking username or password to an AI assistant or agent, saying customers are responsible for activity resulting from shared access. The warning, confirmed by a bank spokesperson to TechCrunch after an email was reported on October 27, comes as companies including Visa and Robinhood discuss AI agents handling payments and trades.
The email, titled “Protecting your finances in the AI World,” says credentials should never be shared with another person or given to an outside service, including an AI-enabled tool or agent acting on a customer’s behalf. Bank of America also said it is “excited about what AI used responsibly can deliver.” A customer posted an image of the email on Bluesky; TechCrunch reported on it, and a bank spokesperson confirmed its authenticity to the outlet.
The message does not describe a new banking product or a specific incident involving a customer’s funds. Its central warning is about who may bear responsibility if an agent uses a customer’s login to take an action. The source article notes that an agent logging in with the customer’s credentials may appear to the bank as the customer—or as someone who has stolen those credentials.
That distinction can matter in either direction: an agent could act without the customer’s intended approval, or a bank could treat its activity as suspicious and restrict access. The source article also cites an October 22 Axios report, based on unnamed sources, that banks were finding third-party AI agents in customer accounts, often using customer passwords. Those claims are attributed to Axios and its sources, rather than confirmed by the bank email itself.
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 Risk of Giving an Agent Full Access
A banking password can grant an agent broad access under the customer’s identity, rather than permission for one limited task. That creates a mismatch: a customer might want an agent to pay a particular bill, while a login could expose more account functions. It may also make it harder for a bank to distinguish a requested action from fraud or an agent’s mistake.
The article raises a further concern about customer liability, but the legal position is not settled for AI agents. It interprets US and EU payment rules as potentially relevant to customers who give software access, while acknowledging that no regulator or court ruling on this specific scenario was identified. Readers should not treat the article’s legal analysis as a definitive determination of their rights.
The practical stakes are broader than a single payment. If an agent can read websites, emails, invoices or documents, misleading instructions embedded in that material could influence its behavior. The source identifies this kind of manipulation, known as prompt injection, as a major AI application security risk. A limited credential or payment approval step could reduce exposure, but does not by itself resolve who is responsible when something goes wrong.
Products Worth Considering
Apple MacBook Pro 16-inch Laptop with M5 Pro, 18-Core CPU, 20-Core GPU, 48GB Memory, 2TB SSD, Space Black
Ledger Flex + Folio Crypto Hardware Wallet
Logitech MX Mechanical Full-Size Wireless Mechanical Keyboard, Tactile Quiet Switches, Backlit Keys, Bluetooth, USB-C, Multi-Device, Graphit
As an affiliate, we earn on qualifying purchases.
Banks Confront Agent Access
The warning lands as financial companies signal that AI agents are already being considered for money-related tasks. At a Tuesday investor discussion, Visa described trust as the biggest barrier to agents paying for goods, according to the source article. Robinhood said customers had placed more than $250 million into accounts that AI agents trade. The source gives no time period or comparison baseline for that figure, so it should be read as a reported amount, not a growth rate.
The source article also describes possible alternatives to sharing a main account password: read-only access, requiring approval for each payment, a virtual card with a spending limit, a token restricted by merchant, amount and time, or a separate account with a limited balance. These are presented as safer design patterns, not as features confirmed to be available from every bank or agent provider.
Robinhood’s reported safeguards do not necessarily settle responsibility. The article says the company has used a separate account and spending limits, while also stating that customers “assume all risk” for trades made by AI agents. That example underscores why customers should read the relevant terms rather than infer that spending controls guarantee protection.
““If you do share your credentials, keep in mind that you’re responsible for activity resulting from that access.””
— Bank of America, in the same email
Liability Rules Remain Untested
No court or regulator ruling applying the cited payment protections to an AI agent was identified in the source material. The article’s reading of US Regulation E—that its exclusion for access furnished by a consumer could be relevant—and its discussion of EU rules are the author’s interpretations, not established findings about how a case would be decided. The source explicitly says the EU question of whether sharing a password with an agent amounts to gross negligence has not been tested in court, as far as the author could determine.
It is also unclear from the available material how each bank would handle a disputed agent transaction, what protections individual AI services provide, or whether a particular bank supports limited permissions or agent-specific credentials. The reported Axios account of banks finding unauthorized agents is based on unnamed sources, and the source does not identify the banks or provide a detailed count. The reported Robinhood amount has no stated measurement window or baseline.
Check Permissions Before Connecting
Customers considering an AI tool for financial tasks should check their bank’s current terms and ask what access the tool receives, how payments are approved, and how disputed activity is handled. Where available, limited permissions, transaction caps and per-payment approval can avoid handing over a main banking password, though no safeguard should be assumed to eliminate financial risk.
Financial institutions and AI providers will need to clarify how agent access is authenticated, how customers can revoke it, and which party handles losses when an agent acts unexpectedly. The source material does not give a timetable for new industry rules or bank products. The next developments to watch are clearer account-access controls, updated terms and any regulator or court decisions addressing responsibility for agent-initiated transactions.
Key Questions
What did Bank of America tell customers about AI agents?
It warned customers not to share banking usernames or passwords with outside services, including AI tools and agents. The bank said customers are responsible for activity resulting from shared access.
Does the warning mean customers automatically lose fraud protection?
Not necessarily. The source discusses possible relevance of US and EU payment rules but says the application to AI agents has not been decided by a regulator or court. A bank’s email is not a definitive ruling on a customer’s legal rights.
What is a safer way to let an AI agent handle a payment?
The source describes options such as read-only access, approval for each payment, a virtual card with a limit, or a token restricted to a merchant, amount and time window. Availability varies, and these measures do not guarantee that losses cannot occur.
What is prompt injection?
Prompt injection is an attempt to steer an AI system through instructions placed in material it reads, such as a webpage, email or document. The source identifies it as a significant security risk for AI applications.
How much money has Robinhood customers reportedly put into AI-traded accounts?
The source article says Robinhood reported more than $250 million in accounts that AI agents trade. It does not give a time period or comparison baseline for that amount.
Source: ThorstenMeyerAI.com
