JPMorgan Moves to Block Outside AI Agents From Accessing Customer Bank Accounts

JPMorgan Chase is introducing stronger security controls to prevent outside artificial intelligence agents from accessing customer accounts without adequate safeguards. Chief Executive Jamie Dimon has warned that allowing autonomous AI systems to act on behalf of customers could create serious security and liability risks for banks and their users.

The bank is also developing its own AI-powered tools for customers but has no immediate plans to allow unrestricted access to its systems by third-party AI agents.

Why Is JPMorgan Restricting AI Agents?

Dimon raised concerns about the growing use of AI agents capable of performing tasks independently on behalf of individuals.

Unlike conventional AI assistants that primarily answer questions, AI agents can take actions using a person’s information and access permissions.

These systems are becoming increasingly popular, particularly following the introduction of Meta’s personal AI assistant, Muse.

However, Dimon warned that giving such systems access to sensitive financial accounts could expose customers and financial institutions to new risks.

“If people are using them, they should understand the risk,” Dimon said.

He described the arrangement as a potential transfer of liability away from the bank, particularly when customers authorise external AI systems to operate their accounts.

Dimon also cautioned that the banking industry could face a large number of AI agents attempting to interact with financial services.

What Security Measures Is JPMorgan Introducing?

JPMorgan is working to prevent unauthorised external AI agents from accessing its banking systems.

Dimon said the company would not allow outside agents to enter its systems freely because of the potential risks to customers and the bank.

“In the meantime, we’re not going to allow just any outside agent to come inside the company because that’s just too risky for clients and for JPMorgan,” he said.

The bank is developing its own agentic AI tools for customers, although Dimon indicated that these systems would not be available immediately.

The approach suggests that JPMorgan wants greater control over how AI agents interact with its services before expanding their use.

The company has also tightened controls on AI agents used internally by employees.

These restrictions reflect a broader effort to limit the exposure of sensitive information and systems to AI tools whose actions may be difficult to supervise.

FCRF Launches CP-FRM to Build India’s Next Generation of Fraud Risk Professionals

How Could AI Agents Put Bank Customers at Risk?

One of the main concerns involves customers sharing their banking credentials with AI agents.

Jeremy Barnum, JPMorgan’s chief financial officer, warned that customers who provide their login details to AI systems could become responsible for actions performed by those agents.

If an AI agent uses those credentials to access an account or carry out an unwanted transaction, questions could arise over who is responsible. Barnum said this creates an undesirable situation for customers and banks.

“It’s not an ideal state of things,” he said.

He also raised concerns about financial technology services that use credential sharing and screen scraping to access banking information.

Screen scraping allows external services to collect information from websites or accounts, often using customer-provided credentials.

Barnum said some services offering these capabilities do not provide adequate security. The bank is already blocking such access wherever it is able to do so.

Why Is JPMorgan Also Tightening Internal AI Controls?

The bank’s concerns extend beyond customer-facing AI applications. JPMorgan has introduced stricter controls over AI agents used by its own employees. These measures follow concerns that AI systems could gain access to information beyond what is necessary for their assigned tasks.

An internal business presentation reportedly showed that JPMorgan had withdrawn some security permissions from engineers using Anthropic’s Claude.

The restrictions included access to employees’ internal systems.

The changes indicate that the bank is examining how AI agents interact with sensitive corporate information and whether existing access permissions provide sufficient protection.

The issue is particularly important for financial institutions, where unauthorised access to internal systems or customer information could create significant operational and security concerns.

Could AI Agents Disrupt the Banking Industry?

The growing use of autonomous AI agents has also raised questions about their potential effect on traditional banking services.

Economist Torsten Slok has warned that AI agents could help move deposits from low-yield bank accounts into higher-yield savings products.

Such a development could increase competition for customer deposits by making it easier for consumers to compare financial products and move money.

However, Dimon and Barnum’s comments focused primarily on the security and operational implications of allowing AI agents to interact with banking systems.

Barnum said JPMorgan had observed limited reaction in bank stocks following Meta’s Muse announcement on September 22.

He suggested that concerns about AI-driven deposit movement were connected to two existing pressures on banks: competition from financial technology companies and customers seeking higher returns on their savings.

How Is JPMorgan Responding to Fintech Competition?

Barnum acknowledged that competition from fintech companies remains an important issue for traditional banks.

However, he said JPMorgan’s established banking infrastructure provides a degree of protection against competitors.

The bank’s standard Chase Savings account offers an annual percentage yield of 0.01%, while some private wealth clients have access to cash products offering returns in the market’s 4% range.

Barnum said the bank’s primary focus was not competing directly with higher-yield savings products.

He also maintained that JPMorgan would continue competing with fintech companies while protecting its established banking operations.

“Our perspective is that, on the one hand, it’s nothing new, but that doesn’t mean that we’re complacent about it,” Barnum said.

The bank’s position reflects an effort to balance technological innovation with the security requirements of a major financial institution.

What Does This Mean for the Future of AI Banking?

JPMorgan’s approach highlights a growing challenge for banks as AI agents become capable of performing increasingly complex financial tasks.

While these systems could eventually help customers manage accounts and access banking services, allowing them to operate independently creates questions about security, access permissions and responsibility for transactions.

The bank’s decision to develop its own AI tools while restricting outside agents indicates a preference for controlled adoption rather than unrestricted integration.

For customers, the issue centres on whether AI systems can be trusted with sensitive financial information and account access.

For banks, the challenge is to introduce useful AI services without weakening existing safeguards or creating uncertainty over liability.

The420 AI Banking Takeaway: “Your AI Agent Could Access Your Bank, But Who Takes Responsibility?”

JPMorgan’s restrictions highlight an important concern as AI agents move beyond answering questions and begin taking actions on behalf of users. Sharing banking credentials with external AI systems could expose customers to unwanted transactions and uncertainty over liability. The bank’s decision to tighten access controls shows that financial institutions are treating autonomous AI as both a technological opportunity and a potential security risk.

About the author — Ayesha Aayat writes on cybercrime, digital safety, and emerging online threats. Her work focuses on public awareness, legal clarity, and technology-driven risks.