Short answer

AI smart glasses put a camera and microphone at eye level, and recordings shared with an AI assistant may be processed or reviewed beyond the device. In March 2026 the UK Information Commissioner's Office said reports about human review of Meta glasses footage were concerning and that it would write to Meta. Wearers should check settings and bystanders should know they have limited control.

Key takeaways

  • Smart glasses can capture photos, video and audio hands-free, so the people around the wearer are recorded too.
  • In March 2026, Swedish newspapers reported that contractors at a Meta subcontractor in Nairobi reviewed footage from Meta's glasses. These are press reports, and Meta says its policies allow human review.
  • The UK ICO called the claims concerning and said it would write to Meta about its obligations under UK data protection law.
  • Meta said data sent to contractors is first filtered to blur faces. Workers quoted in the reports said the filtering sometimes fails.
  • Features that look local, such as live AI, may still send media to cloud services.

Why the glasses change the privacy math

A phone camera is obvious: someone has to lift it. Smart glasses sit on a face, point where the wearer looks and respond to voice commands. When an AI assistant is added, the device can describe what it sees, translate speech or answer questions about the scene. That needs camera and microphone content to be processed, often in the cloud.

The wearer chooses to do that. Everyone in frame did not.

What was reported in March 2026

On 5 March 2026, The Register and Silicon UK reported on an investigation by the Swedish newspapers Svenska Dagbladet and Göteborgs-Posten. According to those reports, workers at a Nairobi-based Meta subcontractor label and review video, audio and transcripts from Meta’s Ray-Ban glasses to help train AI. The workers reportedly described footage ranging from everyday conversations to intimate moments, and said some clips showed bank cards or personal paperwork. Some said the wearer seemed unaware they were recording.

These are claims from the investigation and its sources. Meta’s response, as relayed in the coverage, was that its terms and privacy policy indicate humans may review recorded material, that recordings stay on the device unless the user chooses to share them with Meta AI, and that data sent to contractors is first filtered to blur faces and other identifying details. Workers quoted in the reports said the filtering sometimes fails.

The regulator’s reaction

The UK Information Commissioner’s Office said the claims were “concerning” and that it would write to Meta to request information on how it is meeting its obligations under UK data protection law. It also said providers must clearly explain what data is collected and how it is used. We have not found a published outcome, so treat the matter as open.

What this means in practice

For wearers, the lesson is not that every pair of glasses is spying. It is that “assistant” features can move your recordings, and other people’s, into systems where more people and processes touch them than you expect. Read the data-sharing and human-review sections of the policy, and note that settings can change after updates.

For bystanders, control is limited. You may not know a device is capturing audio, and a small indicator light is easy to miss. Private spaces such as toilets, changing rooms and clinics are where the harm from a recording is greatest, so operators of such places increasingly set no-recording rules.

Wider risks

Recordings can also be used to impersonate people. A few seconds of speech can feed a voice clone, which is why we cover protecting your voice from cloning and AI voice clone scams. Footage of someone’s face can also be misused in abusive ways, covered in deepfake nudes and image-based abuse.

For the general settings that apply to any AI service you use, see AI chatbot privacy settings and data retention and how to opt out of AI training on your data.

A courtesy rule that also protects you

Say when you are recording. Remove the glasses in sensitive places. Keep the AI features off unless you need them. Those three habits cut most of the risk for you and the people around you.

What you can do

  1. 01Before using an AI feature on glasses, read what is sent to the company, who may review it and how long it is kept.
  2. 02Turn off or limit options that share camera or microphone content with the AI assistant when you do not need them.
  3. 03Do not wear recording glasses in private places such as changing rooms, toilets, clinics or other people's homes without clear permission.
  4. 04Tell people when your glasses are recording, and stop if they ask you to.
  5. 05If you are a bystander and believe you were recorded unlawfully or in a private setting, ask the wearer to delete it and raise it with the venue or your data protection authority.
  6. 06Delete recordings and AI interaction history you no longer need.

Related reading: AI voice clone scams and the data breach tracker.

Frequently asked questions

Do smart glasses record all the time?

Not necessarily. Most record when the wearer starts a capture or invokes an AI feature. Some AI features process what the camera sees while active, so check the settings and privacy policy for your model rather than assuming either way.

Can company staff or contractors see what my glasses recorded?

Press reports in March 2026 said contractors reviewed footage from Meta's glasses. Meta told reporters its terms and privacy policy indicate humans may review material, and that data is filtered to blur faces first. Read the policy for your device.

What did the ICO do?

The ICO said the claims were concerning and that it would write to Meta to ask how it is meeting its obligations under UK data protection law. As of the reports we reviewed, no outcome had been published.

What can I do if someone records me with smart glasses?

You can ask them to stop and delete it. In a private setting such as a changing room, tell staff. Rules differ by country, so contact your local data protection authority or police if the recording was covert or abusive.

Sources

  1. UK data watchdog to write to Meta over claims workers viewed intimate footage from smart glasses, The Register, 2026-03-05
  2. Meta contractors review sensitive videos from AI glasses, Silicon UK, 2026-03-05

This article reflects public information as of 2026-11-06. AI tools and rules change quickly; we update articles when new facts are published. Spot an error? See our editorial and corrections policy.