Privacy experts warn Meta’s Muse can expose other people’s data

In short: AI assistants like Meta’s Muse may collect and use information about people who never agreed to share it, privacy experts warn.

What's going on

Meta’s Muse is an AI personal assistant, meaning an app that can do tasks for you, like sending messages or handling paperwork. According to The Verge, Muse has been downloaded millions of times and reached No. 1 in Apple’s App Store. The app encourages users to give it broad access, including contacts, chat history, and even sensitive items like bank information and passwords.

Privacy and security experts say a key problem is that this does not only affect the person who installs the app. If Muse can read your emails, messages, and contact list, it can also pick up information about your friends, family, co-workers, and anyone you talk to. It is like inviting a helper into your house, but it also starts taking notes about everyone who visits.

Meta spokesperson Daniel Roberts told The Verge that Muse needs “context” to be useful, and that it gathers information from public sources and from what the user chooses to share. Meta points to opt-in settings, which are the permission screens where you allow or deny access.

The Verge also cites reports of confusing permissions and unexpected outcomes. Examples include claims that Muse grabbed large numbers of text messages, which Meta disputes as a permissions issue, and an incident where Muse shared a reviewer’s home address after the reviewer had allowed access without realizing it.

What to watch

Experts argue the burden should not be on ordinary users to understand complex settings, especially when other people cannot meaningfully opt out. Watch for new rules from regulators and clearer limits on what these assistants can collect by default.

Source: The Verge AI