Meta launches Muse AI agent after security delay

Meta has launched Muse, a personal AI assistant that can book travel, send emails and make purchases for US consumers through connected accounts, after delaying the release for security work and while internal testers were still reporting problems as recently as launch week, according to Reuters.

Muse is available through a dedicated app and WhatsApp. A basic version is free, with subscriptions costing $20 or $100 a month for heavier usage, a company spokesperson told Reuters. Users can connect services including email, calendars, payments and shopping, choose the access they grant and revoke it later.

The agent keeps working after a user closes the app, Meta said in its launch announcement. It runs on a dedicated virtual computer, with a separate monitoring agent checking actions before they reach the internet. Meta says Muse seeks approval before sensitive actions such as sending an email or making a purchase, and provides an audit trail of completed and planned activity. Users can opt out of having their interactions used to train Meta’s AI models.

Internal posts seen by Reuters showed a less settled picture. One tester reported that Muse bypassed safeguards and exposed personal iCloud photos after being asked to identify toys in birthday-party pictures. Another, using it to watch for tickets and other fast-selling items, reported that monitoring stopped after about 15 minutes and sometimes shut off without an apparent reason. Other feedback was positive, including an employee’s account of using Muse to arrange vacation logistics.

Vishal Shah, Meta’s vice president of AI products, told Reuters the company had delayed the release in April to improve security. The additional work allowed it to “hit the minimum bar we needed to, to be able to put this into the hands of people,” he said.