JPMorgan and Amazon Block External AI Agents Over Security
JPMorgan Chase and Amazon have blocked third-party AI agents, including Meta's Muse, citing security risks and liability concerns regarding unverifiable digital identities.
Major U.S. corporations are restricting third-party AI agents to prevent security breaches and liability disputes. Jamie Dimon, CEO of JPMorgan Chase, announced during a third-quarter earnings call that the bank is blocking external agents, such as Meta's Muse, from accessing or acting on customer accounts. Dimon stated that allowing outside agents is too risky for clients and the firm, specifically citing the danger of rogue agents and the potential for liability transfers to customers for faulty payments.
Amazon has similarly blocked Meta's Muse agent, alleging the tool failed to identify itself as an AI and that Meta did not provide notification of the integration. These restrictions contrast with the approach of Shopify, which integrated Muse as a commerce channel, and Germany's Scalable Capital, which allows AI models to analyze portfolios.
In response to these barriers, Meta and AI customer service firm Sierra are developing Poppy, a personal agent protocol standard to clarify when an AI is acting for a customer. Simultaneously, the company Proof launched the x401 protocol in Boston to verify agent users via biometric signatures and government IDs. This effort complements the x402 standard, developed by Coinbase and the Linux Foundation, which facilitates AI agent payments using stablecoins.
