Overview
- Meta rolled out Muse in the United States on Tuesday, offering a personal AI agent that can open browsers, fill forms, send emails, book travel, negotiate on behalf of users, and continue working in the background.
- Each Muse instance runs inside a dedicated cloud workspace called Muse Secure VM and is overseen by a separate Sentinel monitor that Meta says must approve any internet request before the agent can act.
- Muse can make purchases using Stripe’s Link, which creates one-time-use card numbers and includes purchase protections, and Meta plans two paid tiers reported at about $20 and $100 per month while most uses remain free.
- Internal posts and Reuters reporting show employee tests produced mixed results, including useful task completion, unexplained disconnects, failures to refresh pages, and at least one reported lapse that exposed personal photos, raising concrete reliability and security concerns.
- The launch arrives as Meta tries to reclaim AI leadership, but the product faces trust headwinds from the company’s recent legal and privacy history, including an $18 billion multistate settlement, and Meta says it will offer a stronger Confidential VM later this year to give users more control.