Amazon Blocks Meta Muse AI Agent From Shopping on Its Platform Over Security and Consent Concerns

Amazon began blocking Meta's Muse AI agent from accessing Amazon.com around September 20, 2026, after the company said it never authorized the integration and raised concerns about how the agent handles customer credentials.

Users who tried to use Muse to shop on Amazon were met with an error message stating that access by an unauthorized AI agent violates Amazon's Conditions of Use. Amazon said it had previously asked Meta to voluntarily remove Amazon from the Muse shopping experience, but Meta did not comply.

Amazon cited three main concerns. First, the company said Muse captures and stores customer credentials, which it described as a significant security risk. Second, Amazon said Muse browses without identifying itself as an automated agent, which it called a transparency problem. Third, Amazon argued that third-party applications making purchases on behalf of customers should operate with the merchant's explicit agreement.

Meta responded that Muse does not mishandle sensitive information. The company said the agent does not have direct visibility into user passwords or payment methods, and that credentials shared by users are placed into secure storage that the model can use for transactions without directly reading the data.

The dispute reflects a broader conflict over agentic commerce, in which AI agents make purchases and take actions on behalf of users without direct human input at each step. Amazon has also blocked AI agents from OpenAI, Google, and Perplexity from its platform.

The standoff is notable because Amazon and Meta have an existing commercial relationship. Meta runs AI workloads on Amazon's custom Graviton chips under a multibillion-dollar deal. A security researcher also identified a separate vulnerability in the Muse Mac application that could allow unauthorized local processes to hijack the agent's permissions. Meta issued a hotfix after the disclosure.