Meta is reportedly working on smart glasses that would be recording all the time

2026-07-08 · Source: The Verge · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation, Internet of Things (IoT) & Connected Devices · Depth: Fundamental Awareness, quick

Summary

Meta is reportedly developing "super sensing" smart glasses designed for continuous audio and image recording, capturing what the wearer sees and hears. This prototype, detailed by the Financial Times, would allow users to query Meta AI about the captured data. While raw footage and audio would not be stored or directly accessible to the user, metadata would be extracted and uploaded for AI processing, a design choice Meta proponents suggest has fewer privacy implications. However, the initiative raises significant privacy concerns, particularly as Meta plans to keep the LED recording indicator off during "super sensing" mode, reserving it for "active capture" scenarios. The company is also considering using this continuously captured data for training its AI models and may integrate these features into existing glasses.

Key takeaway

For AI Product Managers developing always-on wearables, you must prioritize transparent privacy controls and clear user consent mechanisms. Meta's approach of disabling LED indicators during "super sensing" mode, while potentially reducing user fatigue, risks significant public backlash and regulatory scrutiny. Ensure your product designs explicitly communicate when data is being captured and how it is used, especially if metadata is extracted for AI training.

Key insights

Meta is prototyping always-on smart glasses that continuously record metadata for AI interaction, raising significant privacy questions.

Principles

• Metadata extraction for AI query aims to mitigate privacy.
• LED indicators are reserved for active user-initiated capture.
• Continuous recording enables "always-on" AI assistants.

Method

The proposed system extracts metadata from continuously recorded audio and images, uploading it to a server for Meta's AI to query, without storing raw footage or making it user-accessible.

In practice

• Integrate "super sensing" features into existing smart glasses.
• Develop AI models using continuously captured environmental data.
• Design privacy controls around metadata, not raw data.

Topics

• Smart Glasses
• Always-on AI
• Privacy Concerns
• Metadata Processing
• AI Model Training
• Wearable Technology

Best for: CTO, Executive, Product Manager, AI Product Manager, AI Ethicist, Tech Journalist

Related on AIssential

• Meta Tapped a Pentagon Supplier to Prototype Face Recognition for Its Glasses — Meta is exploring face recognition for smart glasses using a controversial surveillance tech supplier.
• Meta tests hidden facial recognition code for smart glasses — Meta is internally developing facial recognition for smart glasses, balancing potential utility with significant privacy concerns.
• Meta Is Testing Facial Recognition for Police and Military — Meta is developing facial recognition for smart glasses with a Pentagon supplier, targeting police and military applications.
• Meta allegedly sold AI-enabled smart glasses by leaning hard on “privacy-by-design” marketing—while failing to clearly disclose that... — Wearable AI devices face legal challenges over undisclosed human review of user data and misleading privacy claims.
• We need to admit that putting cameras on AI glasses was a mistake — Camera-equipped AI smart glasses face significant public resistance due to privacy concerns, hindering mass adoption.
• Mind-Reading AI Is Here (No Surgery) — Meta aims for "personal superintelligence" through AI glasses that integrate ambient, non-intrusive AI assistance into daily life.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Verge.