Meta puts its AI assistant on a keychain
Ars Technica Hannah Murphy, Financial Times ● Covered by 10 sources
Meta's AI glasses are getting a camera-free Ray-Ban version and a new assistant. It's betting people will trust wearable AI more than the privacy headaches.
Based on reporting by Ars Technica, Hannah Murphy, Financial Times — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Meta is pushing its AI bet deeper into wearables, even as privacy worries keep circling its glasses. On Wednesday, Mark Zuckerberg unveiled a camera-free pair of Ray-Bans that work with audio only, plus new frame styles across the Ray-Ban line and Meta’s cheaper glasses range.
The company is also adding its Muse assistant to the glasses soon. That matters because Meta has been arguing that portable devices, not just phones or laptops, will be one of the main ways people use AI. Its existing Ray-Bans already include speakers, microphones and cameras, and they can take photos, play music and talk to Meta’s AI.
That hardware has made the product a target. Critics have called the camera-toting glasses “pervert glasses” after intrusive videos spread on Instagram and TikTok, and some say people are using them to record others without consent. Meta does at least put an LED in the frame to show when someone is filming or taking a picture.
The company is also trying to widen the appeal with more than one kind of wearable. It announced new virtual reality glasses for next spring at $1,300, and said they are five times lighter than the previous Quest 3 model. Users will also be able to have a hyper-realistic digital avatar stand in for them on video calls, since the glasses can’t film the wearer’s face.
The bigger plan is more ambitious than the product list. Meta wants Muse to become highly personalized by tapping data from its social platforms, but that runs straight into the usual wall: people will have to connect things like email inboxes, financial services apps and calendars to an AI system that can still make mistakes or go rogue. Meta has already patched a vulnerability that could let a hacker take over another person’s Muse agent, and Amazon has already blocked Meta’s agents from its shopping platform.
My take — AI-written commentary, not fact-checked reporting
Meta keeps selling AI as convenience, but the pitch still sounds like trust-me-first, safeguards-later. That’s a tough order when the product sits on someone’s face, wants their inbox, and has already needed a security fix. The industry loves calling this the future; regular people tend to call it a privacy headache with speakers.
Read more about this at: Ars Technica
