AIWearHQ

Your independent guide to AI wearables.

Reviews, practical buying advice and the latest news on AI Glasses, Smart Rings and AI Earbuds.

AI Glasses | Muse & Privacy

Meta confirms Muse as its glasses’ main AI, as privacy controls improve

Muse is set to become the main AI on Meta glasses, while new privacy controls give owners more choice over visual data. We welcome both developments.

News & analysisBy Published
Four people wearing different styles from Meta’s AI glasses range
Meta’s expanding glasses range. Muse is planned for the glasses; the visual-data setting applies to camera-equipped models. Official image: Meta.

Meta has confirmed that Muse is intended to become the main AI on its glasses, giving owners access to a personal agent that remembers their preferences and can get on with tasks on their behalf. Separately, the company has introduced more choice over whether images and video from visual AI requests are kept and used to improve its models.

Both are welcome developments. Muse could make the glasses considerably more useful, while the privacy change gives people a way to limit an aspect of AI use that has understandably worried some owners and those around them.

Muse will be the AI you talk to on your glasses

In a Financial Express interview published on 29 September, Meta’s VP of AI Products, Vishal Shah, described Muse becoming the “primary agent and AI” on the glasses. The interview took place at Connect, where Meta had already announced the planned integration.

We reported last week that Muse was coming to the glasses. Shah now makes clear that Meta wants wearers to reach their own Muse when they ask for help, with the preferences, goals and connected services they’ve chosen to share. He hasn’t explained every detail of how the switch will work.

There’s an obvious appeal to that continuity. Having to explain the same preferences every time you ask for help soon gets tiresome; an assistant that already knows what you’re trying to achieve ought to need less prompting. How well that works through a pair of glasses is something we’ll want to try for ourselves.

More help with the jobs you keep putting off

Meta’s Connect announcement sets out its ambition to make Muse part of everyday glasses use, from managing a busy calendar to building healthier habits.

Its glasses briefing goes further: proposed uses include booking appointments, checking flight prices, tracking down a missing part and helping buy a product the wearer can see. Meta also describes personalised workouts and meal logging. You’d call your agent by name, give it a job and let it work in the background until it has something to report.

The wider Muse service can already use a browser and connected accounts, according to Meta. Its original announcement says it checks before sensitive actions such as sending an email or making a purchase. Shah explains that permissions start cautiously, with users able to decide how much freedom they’re comfortable giving it; email access can be limited to reading rather than sending.

We like the prospect of handing over some of those small errands while carrying on with our day. We’d still want to check the appointment time or final shopping basket, and Meta’s promised approval steps are reassuring in that respect.

These remain promised capabilities for the glasses, which we haven’t tested. Meta’s announced timetable is the coming months, and the sources above give no confirmed UK launch date. British owners still have a wait before we can judge Muse in everyday use here.

A welcome choice over what happens to visual AI data

Alongside that fuller picture of Muse, Meta has changed the privacy controls on its camera-equipped glasses. A spokesperson confirmed the visual-data opt-out to Business Insider, and the company has published instructions explaining the setting.

Meta distinguishes ordinary photos and videos, which are private by default, from camera data used when asking AI about your surroundings. Turning off visual-data storage means future covered imagery won’t be kept after the request is completed or used for training, product improvement or human review.

Meta lists the control in the Meta AI app under glasses Settings → Meta AI, within Experiences → Store visual data from AI experiences. Switch that setting off to opt out. We haven’t independently checked its availability on a UK account.

There are limits: it doesn’t cover voice interactions, previously processed data, imagery shared through other Meta services, or metadata such as AI-generated descriptions of images. Meta also lists exceptions for content that may breach its terms, including potentially illegal material. Features such as visual reminders may be limited or unavailable.

The opt-out concerns retention and further use; the AI still has to process the request. It isn’t a setting that makes every interaction private or keeps all data on the glasses.

After Hamburg, owners have good reason to welcome this

Our coverage of the Hamburg privacy report highlighted the difference between asking AI for help and allowing people’s information to be used to train it. The regulator’s September assessment raised serious concerns about bystanders who may neither realise what’s happening nor have a meaningful opportunity to object.

For an owner who likes their glasses but feels uneasy about camera data being retained for training, this change could provide some reassurance. It’s perfectly reasonable to want help reading something or identifying an object without wanting the associated images used to improve a commercial AI model.

In our view, giving people that choice is a positive sign that Meta is hearing the concerns surrounding privacy and responding with more control over how data is used. It deserves credit for making the option available and explaining its limits.

We haven’t seen evidence that the change was made specifically in response to Hamburg, and it doesn’t settle the report’s wider questions about recording people and informing them. Owners still need to be considerate about when and where they use the camera.

Our view: more choice is welcome

We’re looking forward to seeing what Muse brings to the glasses. An agent that remembers what we’re working towards and can help finish a job has plenty of appeal, especially when reaching for a phone would interrupt what we’re doing.

We’re also pleased to see Meta giving owners more say over visual AI data. For people who’ve followed the privacy concerns and wondered whether they should use these features at all, a clearly explained choice could make them more comfortable trying them. We’d like Meta to make that choice easy to find and keep explaining what owners are agreeing to as the glasses become more capable.