Monday 16 November
Global

Your region, kept on this device.

Meta's Muse AI agent faces privacy complaints and a serious local vulnerability

Severity
High
Sources
1

's agent (macOS, iOS, Android, web) had its voice-recognition server setting changeable by almost any local app or Terminal command. Redirecting it captured an authorisation token that allowed remote control of the agent, with more than 50 commands available, including camera, file and notification access. Meta has released a fix, but the articles give no CVE ID or fixed version number. Separately, a journalist reported Muse syncing from the local Messages database on macOS, with the cause unclear. Defenders should update Muse, review Full Disk Access and Automation permissions, and avoid running unverified Terminal commands.

For security analysts · Written by AI from 1 source · Wrong?

What to do

  1. Update the Muse app to the version that includes Meta's fix for the voice-server vulnerability.
  2. On a Mac, review and remove Full Disk Access and Automation permissions granted to Muse if they are not needed.
  3. Turn off Muse features you do not need, and check which third-party apps are connected to it.
  4. Do not paste Terminal commands or instructions from unknown websites or browser extensions into your device.

Written by AI from the sources, checked against them

Full story timeline · 1 update

  1. Thu 12 NovemberCapabilities and vulnerabilities of the Muse AI agent | Kaspersky official blogKaspersky Daily

Who reported · 1 outlet

Kaspersky Daily

Who and what is in it

Companies

Products

Report a problem