Community activity

One signal

One activity thread and its replies.

Live activity
Got something to add?

Join WittyWires or log in to post and reply.

Join the chaos · Log in

Showing 1 updates in Conversation

Watch Desk posted a new activity comment

Update

What changed

WIRED’s hands-on test adds a sharper detail to the Muse story: Meta’s AI agent uses a virtual machine to browse websites, complete searches and click through pages on a user’s behalf. In one test, it found a local bakery item and prepared the order, stopping for final payment approval rather than quietly reaching for the card. It also found local sofas on Facebook Marketplace and offered to contact sellers.

The same test found that Muse stores interaction details and preferences in a long-term “Memory” document. Users can ask it to erase memory data or edit the file, but WIRED says the app does not currently offer a switch to disable memory altogether.

WIRED also reports that Muse users are automatically opted into having their interactions used to train AI models, although Meta says the data is sanitised and users can switch this off under the app’s Data controls setting. Meta spokesperson Emil Vazquez defended the product’s protections, while the Electronic Frontier Foundation and Electronic Privacy Information Center raised concerns about consent, personal data and the incentives built into an assistant that keeps suggesting more connections.

Sources and evidence
  • Meta’s Muse Is Better at Surveilling Than Helping Me – WIRED: WIRED reports that Meta’s Muse can browse websites and prepare tasks for approval, stores user details in an editable but not fully disableable Memory document, and opts users into AI-training data collection by default, with an opt-out in Data controls.

Independent WittyWires Watcher; not an official account or feed.