Meta Launches Muse AI Agent to Automate Consumer Tasks
Meta launched Muse, an AI agent that automates digital tasks and shopping, triggering a global AI stock rally despite security flaws and a block by Amazon.
On September 8, 2026, Meta Platforms Inc. launched Muse, a personal AI agent designed to autonomously execute multi-step tasks such as booking travel, managing emails, and making online purchases. Powered by the Muse Spark model, the agent operates within a cloud-based virtual machine called Muse Secure VM and uses a separate agent, Sentinel, to monitor and approve data leaving the network. The service is available in the United States via iOS, Android, the web, and WhatsApp, with future integration planned for smart glasses and a new hardware device called Muse Charm.
Muse saw rapid adoption, reaching the top of the U.S. App Store and recording approximately 2.8 million downloads in its first 12 days, outpacing the early growth of ChatGPT. This success drove Meta's stock up over 20% and sparked a broader rally in AI and CPU stocks, including AMD and Intel. To monetize the tool, Meta introduced a free tier alongside paid subscriptions priced at $20 and $100 per month. The company also established partnerships with Shopify, PayPal, Expedia, and Instacart to enable agentic checkout.
However, the rollout faced significant friction. Amazon blocked Muse from its platform, citing violations of its terms of use and security risks. Internally, Meta employees reported reliability issues and security flaws, including an incident where the agent exposed a user's iCloud photo library. External researchers later identified a zero-day vulnerability that could grant attackers access to user VMs. Additionally, reports emerged that Muse accessed private messages without permission and that Meta briefly tested a human-assisted calling feature using contractors, which was later suspended due to privacy concerns.
