Amazon blocks Meta's Muse AI shopping agent, marking the start of a battle for rules in the era of AI-powered personal shopping.

Amazon blocks Meta's Muse AI shopping agent, marking the start of a battle for rules in the era of AI-powered personal shopping.

AI agents are moving from "helping users search" to "placing orders on behalf of users," but whether e-commerce platforms have the right to refuse these agents from entering their transaction systems is becoming a new point of contention among tech giants.

On September 21, GeekWire reported that Amazon has cut off access to Muse, Meta's personal AI agent, for shopping on Amazon on behalf of users. Users attempting to shop on Amazon through Muse will see a message stating that the unauthorized AI agent's continued access to Amazon violates the terms of use. Amazon stated that it had previously requested Meta to proactively exclude Amazon from Muse's services, but received no response.

This conflict doesn't involve two unrelated companies. Since 2023, Amazon products have been available for direct purchase through Facebook and Instagram; and in April of this year, Meta signed a multi-billion dollar deal with Amazon to run agent AI workloads on Amazon's cloud. Amazon says it is currently in direct communication with Meta, but declined to comment on whether it is considering legal action.

Why did Amazon block Muse?

Amazon's reasons mainly involve authorization, identity disclosure, and account security.

Amazon claims that Meta did not inform Muse beforehand that it would be accessing its platform, and that Muse did not proactively disclose its AI agent identity during browsing. Amazon argues that this amounts to an unauthorized third party entering a user's account, processing transactions, and accessing related data.

Amazon also noted that Muse allows users to access their account page and order history with prompting. An Amazon spokesperson stated that third-party apps offering valet shopping services should operate transparently and respect whether the platform allows their participation.

Meta previously stated that Muse cannot obtain user passwords or payment methods; user-provided credentials are stored securely, and Muse can use related services without directly viewing the credentials.

Muse's product design also touches on this controversy: for services that provide public APIs, Muse can access them using user credentials; if there is no API, users can "use the service through a browser just like a user." Amazon is blocking the latter access method.

Amazon has previously attempted to restrict its AI agents' access to its platform through legal means, but the results of those lawsuits have not been entirely favorable.

Amazon had previously filed a lawsuit against Perplexity's Comet browser and obtained a preliminary injunction in March of this year. However, on August 4, the Ninth Circuit Court of Appeals overturned the injunction, ruling that under federal anti-hacking laws, the person accessing Amazon's computer systems was the user, not the AI company. On September 10, the court further rejected Amazon's application for review.

However, this ruling still leaves room for Amazon to file claims based on contracts and terms of service. The warning issued against Muse did not cite hacking but instead directly addressed Amazon's "Terms of Use," indicating that the focus of the dispute is shifting to whether the AI agent was authorized by the platform.

AI-powered personal shoppers are encroaching on the core interests of e-commerce platforms.

Amazon's tough stance on proxy shopping is also related to its own business model.

Amazon's advertising revenue is projected to exceed $68 billion by 2025. If AI agents directly handle searches, price comparisons, and order placements, users wouldn't need to visit Amazon's website, potentially impacting traditional ad display and product recommendation mechanisms. Over the past year, Amazon has continuously restricted external AI agents' access to its platform; in addition to suing Perplexity, the company has also taken steps to block shopping agents from Google and OpenAI.

Meanwhile, Amazon is also developing its own proxy shopping capabilities. Launched in May of this year, "Alexa for Shopping" includes "Buy for Me," which allows users to search for and purchase goods on external brand websites on their behalf. Amazon emphasizes that this feature proactively discloses the user's identity and allows brands to opt out.

As a result, a core issue arising from AI-powered purchasing agents has gradually emerged: whether platforms should allow third-party AI agents to represent users in their transaction systems, and whether platforms can continue to control transaction entry points and user relationships.

Muse's rapid growth accelerates rule-based game dynamics.

Meta officially launched Muse on September 8, positioning it as a personal AI agent capable of performing multi-step tasks. It can connect to services such as email, calendar, payments, dining, and shopping, and supports iOS, Android, muse.ai, and WhatsApp.

Muse gained attention quickly after its release. Within just one week of its launch, it topped the Apple App Store's US free app chart, surpassing ChatGPT. Early users have already used it to switch car insurance plans, find discount codes, and automatically fill online shopping carts.

Meta stated that Muse runs in a secure virtual machine, is equipped with an independent browser, and obtains user confirmation before sending emails or completing sensitive operations such as shopping. At the same time, the content sent out is reviewed by the Sentinel monitoring agent.

As Muse enters real-world transaction scenarios, the boundaries between AI agents and traditional internet platforms are being pushed to the forefront. Who controls the agent, who owns the user relationships, which services the agent can access, and how transaction value is distributed will all become issues that need to be addressed in the era of "agent commerce."

Risk Warning and DisclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.