Meta debuts its Muse AI agent. Will consumers trust it?
TechCrunch Sarah Perez ● Covered by 4 sources
Meta launched Muse, an AI agent that can handle real tasks like email and bookings. The catch: it needs deep access to your apps, and Meta has a trust problem.
Based on reporting by TechCrunch, Sarah Perez — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Meta is making a very different kind of AI pitch this time. Just under two weeks after agreeing to an $18 billion multistate settlement over social media harms, the company unveiled Muse, a personal AI agent for users in the U.S. This isn’t another chatbot that talks back nicely. Meta wants it to act.
Muse hooks into the services people already live in: email, calendars, payments, health and fitness apps, smart-home tools, dining, shopping, music, events, and more. Meta says it can send emails, book travel, lower bills, fill out forms, build plans, turn recipe reels into grocery lists, send invitations, and even make purchases through Link by Stripe. Shopify’s Shop Pay and 1Password integrations are also on the way.
The company is trying to make that level of access feel less alarming. Users connect apps one at a time, and Muse can use a public API if one exists, or fall back to the browser if it doesn’t. It’s powered by Meta’s Muse Spark model, comes with built-in connectors, and will start on the web, iOS, Android, and WhatsApp before showing up on Meta’s AI glasses.
Money is part of the design from the start. Muse is free to use, but Meta says heavier usage will eventually require a subscription, so a payment card is needed up front. Two paid tiers launch at $20 a month and $100 a month, while Meta says most people will stay on the free plan. The app will show a usage meter and warn users when they’re about to run out.
The bigger problem is not capability. It’s trust. Meta says Muse runs in a dedicated secure computer called Muse Secure VM, with a separate Sentinel agent kept apart at the system level, and that it won’t see passwords, payment methods, or share chat data with ads systems. But Meta’s privacy record is already doing laps around the room, and no amount of tasteful naming fixes that overnight.
My take — AI-written commentary, not fact-checked reporting
Meta keeps trying to sell surveillance-adjacent products as convenience, which is a bold hobby. A personal AI agent that can book, buy, and plan is exactly the sort of thing users may want from a company they don’t have to squint at. Meta is asking for trust first and usefulness second, and that order is usually the problem.
Read more about this at: TechCrunch