JPMorgan Chase has announced that it will block external artificial intelligence agents like Meta's Muse and Spear Street Technology's Instinct from having direct access to customer accounts. CEO Jamie Dimon highlighted the risks posed by these "rogue agents" and emphasized the bank’s commitment to protecting clients and itself from potential threats linked to autonomous AI tools. The bank's decision follows recent security incidents in the AI sector and echoes prior disputes with fintech companies over data access.

While outside AI agents are prevented from executing transactions such as moving funds or paying bills, they may access some customer data in a read-only format through approved intermediaries like Plaid, but only under strict conditions. JPMorgan prioritizes secure customer-authorized data sharing through open banking standards, though the bank has no current plans to permit external agents to interact directly with accounts. This cautious stance aims to mitigate risks associated with unauthorized or malicious use of AI interfaces.

Internally, JPMorgan is actively developing and piloting its own AI applications, such as Smart Cash. This tool forecasts cash flow for users and automatically reallocates funds to optimize yield, positioning the bank to leverage AI benefits while maintaining control over security protocols. CFO Jeremy Barnum detailed that this internal approach reflects the bank’s strategy to use AI safely rather than ban the technology altogether.

Dimon also warned customers about the liability implications of authorizing outside AI agents to act on their behalf. As AI-powered personal financial assistants grow in popularity—demonstrated by recent funding rounds and app download surges—JPMorgan’s firm stance highlights the need for customer awareness regarding potential risks and the transfer of responsibility when allowing third-party AI to manage finances. This protective approach may limit the impact agents have on traditional banking deposits and flows.