When an AI agent logs into a bank account, the bank faces a question it was never designed to answer: Is this the customer, an assistant the customer authorized, or an automated system acting beyond its instructions?
That question is no longer theoretical. Major U.S. banks are blocking some outside AI agents from accessing customer accounts. JPMorgan Chase CEO Jamie Dimon has said the bank is developing tools to manage the risks, warning that unrestricted access by outside agents would be too risky for clients and the bank. Charles Schwab CEO Richard Wurster has said the company is blocking unauthorized agents to maintain control over authentication and the customer experience.
The concern is not simply whether an AI system can log in. It is whether the bank can verify who is directing it, what it is permitted to do and who is responsible if it makes a costly mistake.
The answer may be a new identity system.
The AI Agent Needs An Identity
Today, a bank can recognize a customer through passwords, a phone, a security code, a device or other checks. But when a customer gives an AI agent access to an account by sharing login credentials, the bank may see the customer’s identity even though a machine is acting in the account.
That creates a dangerous gap. A legitimate agent could misunderstand a request, follow malicious instructions hidden in a website, or take an action the customer never intended. A criminal could also disguise an agent or hijack one that already has access.
The Bank for International Settlements has identified authentication, liability and fraud as major challenges for agent-driven payments. The central question shifts from “Is this person allowed to pay?” to “Does this agent have permission to make this specific payment?” Banks will want a reliable way to identify the agent, connect it to its owner and set clear limits on what it can do.
That could mean separate credentials for AI agents, limited permissions, transaction caps, constant monitoring and additional customer verification. Some form of digital identity may become a central part of the answer.
But a system built to verify identity can also become a system that decides whether identity is acceptable.
var images = [],
index = 0;
images[0] = ““;
images[1] = ““;
images[2] = ““;
images[3] = ““;
index = Math.floor(Math.random() * images.length);
document.write(images[index]);
//done
Security Can Become A Gatekeeper
Digital identity is often presented as a convenience: one trusted credential that makes banking, travel, health care and online services easier to access. Biometrics such as a face scan or fingerprint might help confirm that a real person is present. A digital ID could connect that person to an account or authorize an AI agent to act on their behalf.

