Jamie Dimon bars AI agents from JPMorgan accounts: 'You don't want rogue agents'

Jamie Dimon just drew a line through the AI agent gold rush. On Tuesday, the JPMorgan Chase CEO told analysts on the bank's third-quarter earnings call that JPMorgan is blocking outside AI agents from logging in to or acting on customer accounts, citing security risks from 'rogue agents.' '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,' Dimon said. 'You don't want rogue agents.'

The stance carries weight because of what JPMorgan is guarding. The bank reported $15.3 billion in net income for the third quarter, with $1.6 trillion in average loans, $1.5 trillion in cash and marketable securities, and $5.1 trillion in total assets as of Sept. 30. When the largest US bank refuses to let third-party AI agents touch customer accounts, it sets a precedent for how the entire industry treats agentic AI, the wave of tools that can perform tasks on a user's behalf rather than just answer questions. For the tech giants rolling out these agents, banking was supposed to be the killer app; Dimon just made it the hardest sell.

Dimon's warning is not hypothetical. A Cloud Security Alliance research note last month documented a cybercrime campaign in which three open-source AI agents allegedly worked together to breach at least 27 organizations and take more than 600,000 credit card records from two of them. The note urged enterprises to incorporate agentic AI attack scenarios into their threat models for retail and payment infrastructure 'rather than treating agentic AI risk as a generic or future-facing concern, given that this campaign demonstrates the capability is already commoditized and in active use against the retail sector.'

JPMorgan's position, per a spokesperson: 'Protecting our customers is our top priority. When we detect a third-party agent attempting to access accounts by obtaining customers' banking credentials, we do not allow it. We support secure, customer-authorized data sharing through open banking providers.' Dimon added that the bank is building its own tools to protect customers: 'We have to make sure we protect our customer in the right kind of way. So we're building tools to do that, too.'

Not every bank is slamming the door. In August, German bank Scalable Capital opened its platform to AI models like ChatGPT, Claude, and Grok, letting users analyze portfolios, set up savings plans, and place trades through prompts. The guardrails: users had to approve trades, and the assistants could not make payments or withdraw money. Chief product officer Alexander Seipp framed the openness as a competitive edge: 'It certainly creates a level playing field. Access to information and to compute and to intelligence is now available-really-literally in your pocket, 24/7.'

The move complicates the market strategy for AI agents built to manage personal finances. Meta launched Muse in September, pitching it as an agent that can send emails, book travel, and log calendars, but its financial capabilities were a focal point of the rollout. Through an integration with fintech company Plaid, Muse users can connect financial accounts to receive personalized guidance on balances, transactions, and investment holdings. 'The goal is to make financial management feel less like navigating a series of disconnected apps and more like a natural conversation that begins with a person's priorities and is informed by the financial information they choose to share,' a Plaid press release read.

Meta declined to comment on whether Muse could still offer financial assistance if more banks follow JPMorgan's lead. Instead, the company pointed to 'Poppy,' a personal agent protocol standard it is building with Sierra, an AI customer service company, meant to make clear to companies when a personal AI agent is acting for a customer. 'Personal Agent Protocol is complementary,' a Sierra press release read, 'covering how a personal agent identifies itself, gets the customer's permission, and works with a company across any kind of task-from buying a sweater to applying for a mortgage.'

The standoff frames the next phase of agentic AI: the technology's most promising consumer use case, managing money, now runs headfirst into the industry's most security-conscious gatekeepers. Banks that follow JPMorgan will force AI companies to build identity and permission standards like Poppy into their products before they can reach customer accounts. The question for every executive watching is whether the era of rogue agents ends with protocols, or with walls. For now, the safest bet is that the biggest bank in America treats every uninvited AI as a threat until proven otherwise.