Amazon has blocked Meta's Muse AI agent from completing purchases on its site, intensifying a dispute over who controls the online shopping experience when software acts for consumers. By Sunday evening, shoppers attempting to buy through Muse saw a notice saying continued access by an unauthorized AI agent violated Amazon's Conditions of Use. Amazon said it had asked Meta to exclude the site from Muse's scope and argued that third-party purchasing applications should operate openly and respect merchants' decisions about participation. Amazon also said Muse concealed its identity while browsing and appeared to collect and retain customer credentials, potentially exposing account pages and order history without Amazon's knowledge or consent. Meta launched Muse earlier this month for tasks including email, travel bookings, form completion and purchases. The agent operates on a dedicated virtual machine in Meta's cloud, uses Link by Stripe to create single-use card numbers, and is governed by a monitoring system called Sentinel. Muse is available on iOS, Android, the web and WhatsApp, with a free tier and subscriptions costing $20 or $100 per month. Meta says Muse cannot see users' passwords or payment methods and stores shared credentials securely. The companies remain commercial partners through Amazon shopping on Facebook and Instagram since 2023 and an April agreement for Meta to use Amazon's cloud infrastructure for AI workloads. Amazon has previously taken legal action against Perplexity over its Comet browser and targeted shopping agents from Google and OpenAI. Its advertising business generated more than $68 billion last year, tying its financial interest to shoppers browsing product pages and sponsored listings. Amazon said it was speaking directly with Meta and declined to say whether it would pursue legal action.