UK and EU Probe Meta Over Private Smart Glasses Videos

Regulators in the United Kingdom and the European Union have begun examining Meta after reports suggested that contractors reviewing data from Ray-Ban Meta AI smart glasses have been exposed to highly intimate footage recorded by users. The investigation raises questions about how user data from wearable AI devices is collected, reviewed, and potentially used for training artificial intelligence systems.

The UK’s Information Commissioner’s Office (ICO) confirmed it has contacted Meta to seek clarification following media reports that outsourced workers in Kenya are reviewing video captured by the smart glasses. The regulator said the allegations are concerning, particularly regarding whether people who appear in the recordings are aware that human reviewers may see the footage.

Investigation Triggered by Media Reports

The regulatory scrutiny began after a Swedish investigation by newspapers Svenska Dagbladet and Göteborgs-Posten revealed details of the content moderation process behind Meta’s smart glasses.

According to the report, contractors hired by Meta through outsourcing partners review video clips captured by Ray-Ban Meta AI glasses in order to label and categorize the material for machine learning systems.

These reviews are meant to help train AI models and improve features such as object recognition, voice commands, and contextual assistance.

However, the investigation claimed that some of the footage being reviewed includes extremely private situations recorded unintentionally by users.

The ICO said it is now making enquiries with Meta to determine whether the company’s data handling practices comply with UK data protection rules.

Workers Describe Viewing Highly Private Footage

Data annotators working for the Kenyan contractor Sama told reporters they have been required to watch and categorize deeply personal recordings captured by the glasses.

According to the investigation, some of the material includes people undressing in bedrooms, individuals using toilets, and couples engaged in sexual activity. Workers also reported seeing screens and payment cards that revealed sensitive financial information.

In some cases, the recordings appeared accidental. For example, workers described videos where someone left their smart glasses on a table while the camera continued recording activity in a private room.

Because of these situations, annotators said they often felt that the people being filmed were unlikely to know that the footage might be reviewed by human workers during the AI training process.

How Meta Explains the Data Review Process

Meta acknowledged that certain interactions and recordings from Ray-Ban Meta smart glasses can be reviewed by human annotators.

The company says this process is used to improve the product and train AI systems that power the device’s capabilities. According to Meta, the possibility of human review is disclosed in its privacy policy and AI terms of service.

Meta also says it encourages users not to capture sensitive information when using the glasses.

The company states that several safeguards are in place to reduce exposure to sensitive material. These measures include filtering systems that detect inappropriate content, blurring technology designed to obscure personal details, and sampling methods that limit how much data is reviewed.

However, Meta has acknowledged that these safeguards are not perfect and that sensitive footage can sometimes reach human reviewers.

Privacy Concerns Under European Law

The situation has drawn particular attention from European privacy experts because the reported data review process involves sending footage to annotators outside the European Union.

Under the EU’s General Data Protection Regulation (GDPR), companies must meet strict requirements when transferring personal data across borders.

Experts say the reported practice could raise questions related to transparency, data minimization, and consent. One concern is that individuals appearing in the videos may not know they are being recorded by smart glasses, let alone that the footage could be reviewed by workers in another country.

Critics argue that simply including such practices in lengthy terms of service documents may not satisfy the requirement for meaningful consent, especially when bystanders are involved.

Previous Criticism of Smart Glasses Technology

Meta’s Ray-Ban smart glasses have faced privacy concerns since their launch.

The wearable device includes a built-in camera, microphone, and AI features that allow users to capture photos and videos, livestream content, and ask the AI assistant questions about what it sees.

Privacy advocates have long warned that wearable cameras could create situations where people are recorded without realizing it.

Some critics have also raised concerns about the possibility of future features such as facial recognition or continuous recording that could build a detailed digital memory of a user’s daily activities.

Although Meta says the glasses include indicator lights to signal when recording is happening, privacy groups argue that these signals can be easy to miss in crowded environments.

Role of the Kenyan Contractor

The contractor involved in the data labeling work is Sama, a company that provides AI training data services to major technology firms.

Sama has previously been involved in controversy over its role in moderating content for Facebook and Instagram. Former workers filed lawsuits alleging exposure to graphic material and difficult working conditions.

The new reports about smart glasses footage have renewed concerns about the types of content workers are asked to review as part of AI training pipelines.

Worker advocacy groups have argued that companies using outsourced moderators should provide stronger protections and clearer guidelines around sensitive material.

What Regulators May Do Next

The UK Information Commissioner’s Office said it is assessing whether Meta’s disclosures and safeguards meet the requirements of UK data protection laws.

Because the original investigation involved European users, regulators in the European Union may also examine the issue.

Possible regulatory actions could include requiring clearer warnings on smart glasses devices, limiting what types of recordings can be used for AI training, or imposing stricter controls on how sensitive footage is anonymized before it is reviewed.

In more serious cases, regulators could issue financial penalties or restrict cross-border transfers of certain types of personal data.

Growing Scrutiny of AI Data Practices

The controversy highlights a broader challenge facing technology companies as they develop AI products that rely on large volumes of real-world data.

Many AI systems require human reviewers to label or verify information so that algorithms can learn from it. However, when that data includes real recordings from people’s homes and daily lives, the ethical and legal implications become more complicated.

As wearable devices powered by AI become more common, regulators are likely to examine more closely how companies collect, process, and store the data these devices generate.

For Meta, the investigation into Ray-Ban smart glasses may become an important test of how privacy rules apply to the next generation of AI hardware.