Meta debuts its Muse AI agent. Will consumers trust it?
Meta has unveiled Muse, a personal AI agent that connects to users’ email, calendars, payments and other everyday apps to perform tasks like booking travel, sending emails and making purchases. The service, powered by Meta’s Muse Spark model, will launch in the U.S. across web, iOS, Android and WhatsApp with a free tier and two paid plans at $20 and $100 per month.
Why It Matters
Muse asks users to grant deep access to sensitive services and data at a time when Meta faces heightened regulatory scrutiny and lingering trust concerns after recent settlements and past privacy scandals. Whether consumers accept handing over that level of access will be a central test of Meta’s ability to rebuild trust while moving from chatbots to agents that act on users’ behalf.
Key Facts
- Product name: Muse
- AI model: Muse Spark
- Launch region: United States (initial)
- Availability channels: Web (muse.ai), iOS and Android apps, WhatsApp chats; coming to Meta's AI glasses soon
- Core capabilities: Send emails, book travel, lower bills, fill out forms, create plans, convert recipe reels to grocery lists, send party invitations, make purchases via Link by Stripe (Shop Pay and 1Password integrations coming)
Meta on Tuesday introduced Muse, a personal AI agent designed to connect to the services people use every day and carry out tasks for them. The company says Muse can link to email, calendars, payment systems, health and fitness apps, smart-home services, shopping and dining apps, music and events platforms, and more. When a public API is available Meta will use it; otherwise Muse can set up connections using user-provided credentials or access services through a browser.
The agent is intended to move beyond question-and-answer chatbots by performing actions on users’ behalf — from booking travel and sending emails to lowering bills and making purchases. Meta plans to leverage Link by Stripe for checkout, which the company says offers purchase protections; Shopify’s Shop Pay and 1Password integrations were also announced as forthcoming. Muse will be free to start, but requires a payment card to enroll and includes two paid tiers at launch: Power for $20 per month and Maximum for $100 per month. Meta says most users will remain on the free tier and that the app shows a usage meter and will warn when free usage is exhausted.
On the technical and privacy side, Meta says Muse runs inside a dedicated virtual machine called Muse Secure VM and that a separate Sentinel agent runs on the same VM but is isolated at the system level. The company claims Muse will not have visibility into users’ passwords or payment methods and that conversations or user data will not be used by Meta’s advertising systems. Muse also learns from users’ conversations to provide ongoing, unprompted suggestions and can continue working after users leave the app.
The product arrives as other firms and startups push agent-style AI that performs tasks for users, and amid renewed scrutiny of Meta’s privacy record. The announcement came less than two weeks after Meta agreed to an $18 billion multistate settlement in a lawsuit over consumer harms. Meta previously faced enforcement actions and settlements — including a 2019 $5 billion FTC settlement and a 2011 FTC matter — and incidents such as exposed readable passwords in 2019 and the Cambridge Analytica data scandal. Meta’s technical claims about Muse’s protections are explained in a separate technical post, which the company says will require deeper review by security experts.
Keep Reading
Razer’s ANC-equipped gaming headset for PS5 and Xbox is almost $100 off
Rivian spinout Also apologizes for delays in shipping futuristic e-bikes
Drama swirls around OpenAI’s legendary mathematical milestone
