Amazon says Muse AI does not identify itself during site browsing, and alleged it is capturing customer logins|Malkawi99|CC BY-SA 4.0
Amazon has barred Meta’s AI assistant, Muse, from accessing its e-commerce site, saying that access by an unauthorized AI agent violates the conditions of use customers have agreed to.
According to Amazon, the app was never given permission to browse its site and claims Muse creates privacy and security risks.
Meta introduced the AI agent earlier this month to let users automate tasks such as filling out forms, sending emails, shopping, and managing personal calendars.
However, Amazon says the personal assistant does not identify itself during site browsing and alleges it captures customer logins. Meta countered that login data remains securely stored without direct agent visibility.
Amazon’s block cuts off Muse’s access to the largest online store in the US. This may affect how many people use the AI agent and could also redirect shopping traffic to other e-commerce sites that allow it.
At the same time, Amazon’s products have been listed on Facebook and Instagram since 2023.
This isn’t the first time Amazon has kept rival agentic apps out of its platform. Last year, it sued Perplexity’s Comet browser agent, saying the AI assistant accessed the site without permission and posed risks to Amazon’s security systems. A federal appeals court said a shopper directs the AI and hence the user is directly involved in the purchase. It overturned a previous order barring Comet. The case is still in court.
AI agents are a high-stakes race for Amazon. CEO Andy Jassy revealed in July that over 350 million shoppers used its internal shopping assistant, Alexa for Shopping, last year. He added that, on average, US customers using the tool spent 40% more per order.