TL;DR

Meta’s latest smart glasses include on-device facial recognition components, including detection and embedding models. However, it is unclear whether these features are active for users. The technology’s presence raises privacy and security questions.

Meta’s smart glasses contain fully assembled facial recognition hardware and software, but it is not yet clear whether these features are active for users. The discovery was made through an inspection of the Stella app’s code, revealing the presence of detection models, biometric embedding systems, and a local face database. This suggests Meta has the technical capability to perform facial recognition directly on the device, raising privacy implications.

Researcher analyzing version 273.0.0.21 of Meta’s Stella app identified the complete set of on-device facial recognition components, including three face models, a biometric embedding generator, and a local database schema designed for storing and matching facial data. The models, based on open-source architectures, are stored locally on the glasses and capable of detecting faces, extracting biometric fingerprints, and performing similarity searches within a vector index.

Tests showed the recognition pipeline can detect a face, generate a 2048-dimensional biometric embedding, and trigger a notification if a match is found. However, the researcher noted that on a standard, unenrolled user account, the user interface for recognition is not active, and no identity data was pushed to Meta’s servers during testing. The hardware and software components are present and functional but appear gated behind user enrollment or specific conditions.

Why It Matters

This discovery indicates Meta has the technical infrastructure for facial recognition directly on its smart glasses, which could enable future features such as real-time identification of people in view. The presence of this capability raises privacy concerns, especially regarding user consent and data security, as the system could potentially identify individuals without explicit approval. It also signals a possible shift toward more invasive augmented reality functionalities.

Lockin Venoris Solar Face Smart Lock – Facial Recognition Keyless Entry Deadbolt- Solar Powered for Perpetual Battery Life – Built-in WiFi- Works with Apple Home, Alexa & Google Home- No Subscription

Lockin Venoris Solar Face Smart Lock - Facial Recognition Keyless Entry Deadbolt- Solar Powered for Perpetual Battery Life - Built-in WiFi- Works with Apple Home, Alexa & Google Home- No Subscription

Smart lock with facial recognition, solar power, and smart home integration for secure, keyless entry.

Power SourcePerovskite Solar & USB-C backup
Recognition Method3D Facial Recognition
ConnectivityBuilt-in WiFi & Smart Home
Unlock MethodsFacial, PIN, App, Voice, Mechanical Key

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Background

Meta has been developing smart glasses for several years, with previous models emphasizing augmented reality and social features. The recent code inspection suggests that the company has integrated facial recognition models into the hardware, similar to those used in mobile apps and academic projects, but has not yet activated these features for general users. Past reports have indicated Meta’s interest in biometric data, but there has been no official confirmation of active facial recognition in consumer devices until now.

“The machinery is present, it is wired together, and capable of recognizing faces locally on the device.”

— Researcher analyzing the Stella app

“N/A”

— Meta spokesperson (not yet available)

What Remains Unclear

It remains unclear whether Meta plans to activate facial recognition for users, under what conditions, or how they will handle privacy and consent. The current code suggests the capability exists but is not yet operational for the general user base. Details about future deployment, data handling policies, or user controls are not yet available.

What’s Next

Meta may choose to activate these features in future updates or hardware releases, possibly accompanied by privacy policies or user controls. Observers will likely monitor updates to the Stella app and device firmware for any signs of activation or official announcements.

Key Questions

Is Meta currently using facial recognition on its smart glasses?

Based on current code analysis, the hardware and software for facial recognition are present and functional, but it is not confirmed that the feature is active for users at this time.

It is not yet clear whether Meta plans to activate these features broadly or with user consent. The current status suggests gating or testing controls are in place.

What are the privacy implications of this technology?

If activated, facial recognition could enable real-time identification without explicit consent, raising significant privacy and security concerns. How Meta will address these issues remains to be seen.

When might this feature be activated for users?

There is no official timeline. Future updates or hardware releases could include activation, depending on Meta’s policies and regulatory considerations.

Will users be able to disable facial recognition?

It is unknown whether controls will be provided. Future policies and user interface options will determine if and how users can opt out.

Source: Hacker News

You May Also Like

How To Survive An Attack By Robotic Dogs

Avoid becoming a target by learning essential survival tactics against robotic dog attacks—your safety may depend on it.

Spy Vs Sensor: Tech Tricks Agents Use to Avoid Detection

Spy vs Sensor: Tech tricks agents use to avoid detection reveal layered strategies that keep them hidden from ever-evolving surveillance systems—discover the secrets behind their stealth.

Dark Web Monitoring: Tech That Hunts Threats in Hidden Corners of the Internet

Only by understanding dark web threats can you truly protect your assets from unseen dangers lurking beneath the surface.

Wirestock raises $23M to supply creative multi-modal data to AI labs

Wirestock raises $23 million to provide AI research labs with diverse multi-modal data, including images, videos, and audio, to support AI training.