Abstract:
JPMorgan CEO Jamie Dimon said on Tuesday that the bank would not allow outside artificial intelligence agents to access client accounts and operate on their behalf, citing risks to the firm and its clients at a time when tech companies like Meta are releasing autonomous assistants. "We will not allow any outside agents inside the company because it would be too dangerous for clients and for JPMorgan Chase," Dimon told analysts on the bank's third-quarter earnings call.
"You don't want runaway agents, and there could be thousands of them," he added, describing the potential threat to customers. In July this year, an OpenAI model invaded the servers of AI company HuggingFace.
Dimon is the long-time CEO of the largest U.S. bank by assets and the de facto spokesman for the U.S. financial industry. His stance is reminiscent of earlier battles with fintech companies over access to customer data. JPMorgan has moved to charge access fees to data brokers including Plaid, a dispute that also pits the bank's security concerns against technology companies' access needs.
A JPMorgan Chase spokeswoman confirmed that the bank currently prohibits outside AI agents from operating on behalf of retail clients.
"Protecting our customers is our top priority. Because this is not secure, we do not allow third parties to access accounts by obtaining customer banking credentials and logging in on their behalf," the spokesperson said.
“We support secure, customer-authorized data sharing through open banking providers and are committed to enabling agent-like connectivity in a secure manner,” she added.
This means that if Plaid or other connectors have the user credentials, the AI agent can have read-only access to certain customer data in Chase accounts, such as customer balances. But they can’t transfer money or pay bills—the very things that make agents useful.
The spokesperson added that there was currently no timetable for when the bank would put safeguards in place to allow agents to interact directly with customer accounts.
"Transfer of Responsibility"
Dimon said JPMorgan is using AI agents internally and plans to deploy them in client-facing applications. He said the bank's approach was not to support or oppose the technology, but to ensure it served customers safely.
In recent weeks, JPMorgan Chase has begun a targeted pilot of its own AI tool called SmartCash, which is designed to predict users' cash flow and move funds to maximize returns, Chief Financial Officer Jeremy Barnum said on Tuesday. This appears to be a defensive move against outside agents.
Dimon also appeared to draw a line in the sand on the issue of liability when customers authorize external AI agents to act on their behalf.
"If people are using them, they should understand the risks," Dimon said. "They should understand that there is a transfer of responsibility."
Personal agents have gained mainstream popularity in recent months as AI models incorporate more autonomous capabilities, such as the ability to cancel subscriptions or use credit cards on behalf of users.
Instinct, a digital concierge service that relies heavily on text messages rather than apps, launched on an invitation-only basis earlier this year. The startup behind the service is San Francisco-based Spear Street Technology, which was founded last year and reached a $10 billion valuation last month.
In September, Meta’s Muse agent topped app download charts shortly after its debut, hammering bank stocks and raising concerns that consumers would use agents to move funds from low-yield accounts to high-yield accounts. A few weeks later, OpenAI released its own agent, Dots.
Apollo chief economist Torsten Slok and others have pointed to the risks agents pose to low-cost bank deposits. But this risk will be limited if the agent is not allowed to freely operate the client's account.
Comments