Dimon warns JPM customers about AI agents and “liability transfer” risks

JPMorgan Chase CEO Jamie Dimon said customers using artificial intelligence agents need to understand a specific category of risk that can emerge when decisions and communications are automated. In remarks reported in a market news article Tuesday, Dimon warned about what he described as “liability transfer” in the context of AI, urging customers to consider what happens when an AI system triggers actions and the human or institution on the other side may face unclear responsibility.

The comments landed as renewed discussion spread in financial circles about the possibility of “deposit flight” driven by AI. The same report ties the latest wave of concern to Meta’s Muse, an AI assistant featured in recent coverage that has re-ignited broader debate about whether AI interfaces could make it easier for customers to shift balances among institutions.

Dimon’s core message, as described in the report, was not that banks will be unable to manage technology risk, but that customers should recognize how accountability can become muddled when AI agents initiate interactions on a person’s behalf. The warning points to a practical problem in banking operations and compliance: when automation takes actions that have financial consequences, the relevant party for approvals, verification, and follow-through may not be as straightforward as with traditional customer workflows.

While the report frames the issue around liability and responsibility, it does not detail the exact mechanisms Dimon was referring to, such as which AI agent actions most commonly create the “liability transfer” scenario. It also does not specify whether JPMorgan has issued any new customer guidance, product terms, or system-level controls in response to the concern.

Meta’s Muse, referenced in the report, is discussed in the broader context of AI assistants that can help users execute tasks through natural language prompts. In banking, that convenience can raise questions about who is accountable if an AI system persuades, advises, or acts on a user’s behalf in a way that leads to a move of deposits, an account change, or a disputed transaction.

Industry concern about AI-driven deposit shifts is tied to the idea that banks could face faster and more frequent balance reallocation if customers can be guided, through conversational interfaces, toward alternative providers. If AI tooling reduces friction in comparing offers and moving funds, the market could become more dynamic, increasing the value of customer retention, onboarding safeguards, and dispute-handling readiness.

JPMorgan’s disclosure in the report appears limited to Dimon’s warning about liability and customer awareness. The article, as summarized, does not provide additional specifics on JPMorgan’s internal risk assessment, whether the bank has changed contract language, or how it is testing AI-related scenarios across customer channels.

For now, the clearest takeaway is that leading banks are beginning to treat AI agents as a governance and accountability issue, not just a customer-experience upgrade. What remains unclear is how regulators and the banks themselves will define responsibility in edge cases, and what standardized controls or disclosures may emerge to make “liability transfer” scenarios easier to evaluate.