Amazon has blocked Meta's AI agent, Muse, from accessing its shopping platform, preventing the tool from browsing products or placing orders on behalf of users. Amazon stated that Muse failed to identify itself as an AI agent, which violates the platform's terms of service regarding unauthorized third-party access to user accounts and sensitive data.
The conflict arose after Amazon's request to exclude its site from Muse's shopping capabilities was rejected by Meta. Amazon expressed security concerns regarding how Muse handles user login credentials, noting that the agent operates by mimicking human browsing behavior without explicit authorization or transparency, unlike traditional third-party services.
Meta, which launched Muse on September 8, maintains that the agent operates within a secure virtual machine environment and does not directly expose user passwords or payment information. The company claims that sensitive actions require user confirmation and are monitored by an independent system called Sentinel.
This move is part of Amazon's broader strategy to restrict external AI shopping agents, following similar actions against Perplexity, Google, and OpenAI. While Amazon previously sued Perplexity over unauthorized access, recent court rulings have shifted the focus toward contract law and platform terms of service as the primary legal battleground.
Despite the current friction, the two companies maintain a commercial relationship; Meta uses Amazon's cloud infrastructure for AI workloads, and Amazon products are available for purchase via Facebook and Instagram. Negotiations between the companies are ongoing as they navigate the evolving landscape of agent-based commerce.
