🔍 Read the full analysis: Your AI Agent And Your Bank Password: Draw The Line Here on ThorstenMeyerAI.com
Get the latest gadgets 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 give banking usernames or passwords to AI tools, saying customers are responsible for activity resulting from shared access. The warning comes as financial firms report growing use of AI agents around payments and trading, while the protections and liability rules for agent-driven actions remain unclear.
Bank of America has warned customers not to share their banking usernames or passwords with AI agents, saying customers are responsible for activity resulting from that access. The email, confirmed as authentic by a bank spokesperson to TechCrunch, arrives as financial companies report rising use of AI agents in payments and trading—and as the rules for disputed transactions involving those agents remain unsettled.
The email, titled “Protecting your finances in the AI World,” says customers should never share their credentials with another person or provide them to an outside service, including an AI-enabled tool or agent acting on their behalf. It adds that customers who do share credentials are responsible for activity resulting from that access. The bank also said it is “excited about what AI used responsibly can deliver.” A customer posted the email on Bluesky, and TechCrunch reported on it October 27. A Bank of America spokesperson confirmed the email was authentic to TechCrunch.
The warning was reported amid signs that AI agents are already interacting with financial accounts. The source article says Visa told investors on Tuesday that trust is the biggest barrier to AI agents that make payments. It also says Robinhood reported customers had placed more than $250 million into accounts that AI agents trade. Robinhood’s stated safeguards include separate accounts and spending limits, but the company says customers “assume all risk” for trades made by AI agents.
Separately, Axios reported October 22, citing unnamed sources, that major U.S. banks were finding unauthorized third-party AI agents inside customer accounts, generally using customers’ own passwords. According to that report, banks often could not tell whether the activity had been requested by the customer. These claims come from the cited reporting; the source material does not identify the banks or provide public data on the number of affected accounts.
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 Agents Full Access
A bank password can grant an agent access well beyond a single task. When an agent logs in using a customer’s credentials, the bank may see activity that looks like the customer’s own—or like a fraudster who has stolen the password. That can leave customers facing unwanted transactions, account restrictions, or a dispute over whether they authorized the action.
Who bears the financial loss is not settled. The source article interprets U.S. Regulation E as potentially excluding transfers initiated by a person given access to an account by the consumer. It also notes that EU payment rules require customers to take reasonable steps to protect their security credentials. But whether an AI agent, or the company operating it, falls under those rules in a particular case has not been decided by a court or regulator, according to the source. The legal discussion is the author’s interpretation, not legal advice.
There is also a security concern: agents process web pages, emails, invoices and other material that may contain instructions intended to manipulate them. The source identifies this as prompt injection and says OWASP ranks it as a leading risk for AI applications. If an agent can access a bank account with a customer’s full login, a misleading instruction could have consequences beyond the agent’s original task. That is a risk to manage, not proof that a particular agent has been manipulated.
Products Worth Considering
Logitech MX Master 4 for Mac, Ergonomic Bluetooth Mouse with Haptic Feedback, Ultra-Fast Scrolling, USB-C Charging, for MacOS, iPadOS, Windo
Apple iPad Air (2025) 13‑inch Wi‑Fi 256 GB – Space Gray (M3 chip)
Samsung 27-inch Odyssey OLED G6 QD-OLED QHD Gaming Monitor
As an affiliate, we earn on qualifying purchases.
Banks Face Unclear Agent Activity
The warning follows a broader shift from AI systems that answer questions to agents that can take actions on a user’s behalf. Those actions may include interacting with shopping services, financial accounts or trading tools. A customer may intend to authorize one narrow task, while a password-based login can provide access as broad as the customer’s own.
The source article contrasts password sharing with more limited forms of access: read-only permissions, requiring approval for each payment, virtual cards with spending limits, or tokens restricted by merchant, amount and time. A separate account with a limited balance is another possible safeguard. These approaches can reduce what an agent is able to do, but technical limits do not automatically settle liability if a transaction is disputed.
Financial companies’ responses are also taking shape. The reported bank activity and Robinhood’s separate-account approach point to different ways firms are trying to manage agent access. The available material does not establish a common industry standard or show that every bank supports scoped access for AI agents.
“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 the customer email confirmed to TechCrunch
Liability Rules Are Untested
No court or regulator ruling on AI agents and shared banking credentials is identified in the source material. The application of U.S. Regulation E to an AI agent or its operator is presented as the author’s reading, while the EU question of whether sharing credentials amounts to gross negligence is also described as untested. Outcomes may depend on the transaction, the access granted, the customer’s actions and the relevant jurisdiction.
It is also unclear how widespread unauthorized agent access is. The Axios report cited unnamed sources, and the material supplies no account totals, named institutions or details of specific losses. The reported Robinhood figure is a total amount customers placed in accounts that AI agents trade; it does not establish how many customers or transactions are involved, or whether those trades produced gains or losses.
The source does not specify whether Bank of America’s email changes existing account terms, what remedies remain available after an agent-related transaction, or what technical safeguards the bank would support. The bank’s warning should not be read as a legal ruling that customers always bear every loss.
Watch for Safer Access Controls
Customers considering financial agents should check the permissions an app requests and whether they can use limited access instead of entering a bank password. Options to ask about include read-only connections, approval for each payment, capped virtual cards, merchant- and time-limited tokens, or a separate account holding only a limited balance. They should also review the service’s terms to understand how it describes responsibility for actions taken by an agent.
For banks and payment providers, the next developments to watch are clearer ways to identify agent activity, define customer authorization and offer restricted credentials. The source material gives no timetable for new rules or industry standards. Until courts, regulators or firms provide more specific guidance, the scope of access and the terms governing losses remain central questions.
Key Questions
What did Bank of America tell customers?
It warned customers not to share their usernames or passwords with other people or outside services, including AI agents. The email says customers are responsible for activity resulting from credentials they share.
Does the warning mean customers always pay for an AI agent’s mistakes?
No. The email states the bank’s position about shared credentials, but the source material identifies no court or regulator ruling that settles how existing consumer-protection rules apply to AI agents. A particular dispute may depend on the facts and jurisdiction.
What can customers use instead of giving an agent a bank password?
Possible alternatives include read-only access, approval for each payment, a virtual card with a spending cap, a token limited to a merchant and time period, or a separate account with a limited balance. Availability varies by service.
What is prompt injection?
It is an attempt to steer an AI system through instructions embedded in material it reads, such as a web page, email or file. The source identifies it as a risk for AI applications; it does not report that a specific bank agent was compromised this way.
How widespread is AI-agent use in financial accounts?
The source reports that Robinhood customers had placed more than $250 million into accounts traded by AI agents and cites an Axios report of banks finding unauthorized agents in customer accounts. It provides no overall count of users or affected accounts, so the scale is unclear.
Source: ThorstenMeyerAI.com
