UK regulator contacts Meta over AI glasses video concerns

The Information Commissioner’s Office (ICO), said it will write to Meta after Swedish papers reported that Nairobi-based workers could view sensitive footage from Ray‑Ban Meta glasses.

Svenska Dagbladet and Goteborgs‑Posten reported workers viewed videos showing glasses-wearers using the toilet and engaging in sex.

The papers said the workers, who labelled images and checked AI transcripts, were data annotators employed by Nairobi-based Sama.

Meta said subcontractors sometimes review images and videos to improve the user experience and that data is filtered to protect privacy, though sources who spoke to the papers said blurring sometimes failed.

The UK data watchdog said devices processing personal data should put users in control and provide transparency.

It said it would request information from Meta.

🔗 Source: BBC

🧠 Food for thought

Implications, context, and why it matters.

Weak safeguards and reported plans for more surveillance deepen Meta’s issues

Some users may not realize that human contractors sometimes review videos and images, with one data reviewer quoted as saying, “I do not think they know, because if they knew, they would not be recording” 1.
Meta’s main safeguard is a small LED recording light on the glasses, yet it can be easy to miss and it can reportedly be disabled without authorization 2.
Meta says its filtering can blur faces, though sources who spoke to Swedish papers plus workers involved in the process report that blurring sometimes fails, including in low-light conditions, which can leave people recognizable to reviewers 3.
Concerns rise after an internal Meta memo surfaced that says the company plans to add facial recognition, which privacy advocates warn could raise risks like stalking and harassment 2.

Smart glasses push courts and the public to rethink privacy in shared spaces

This incident fits a wider backlash as some U.S. courts and court systems have moved to ban smart glasses from courthouses, and judges can order people to remove them, over worries about covert recording plus threats to jurors and witnesses 2.
The devices blur the line between a personal accessory and a tool for watching others, which strains social expectations around consent in public places 4.
The controversy puts pressure on the tech industry’s practice of burying data-use permissions in long privacy policies and terms of service, raising doubts about whether users grasp that human review can happen 3.
As AI wearables spread, this case may shape how clearly companies must spell out human-in-the-loop (human-reviewed) data processing pipelines 4.