According to the Financial Times, Meta is developing a prototype smart glasses with round-the-clock multimodal sensing capabilities. The device supports a "super perception" mode, which can continuously record audio and take photos at intervals of a few seconds. Users can ask immediate questions and interact with the captured audio and video content through Meta AI. To alleviate concerns about data privacy, Meta plans to adopt a unique metadata processing mechanism: the system will not store original videos and audio, nor will it be directly accessible to users. Instead, it will extract metadata and upload it to the server for AI queries.
However, this high-frequency capture feature has still sparked significant privacy controversy. Meta has previously faced strict scrutiny due to facial recognition and eavesdropping incidents. In response to black market activities where users intentionally block the recording indicator light, Meta announced on Tuesday that it would release a system update. If the LED light is tampered with, the camera will be disabled directly. However, according to the white paper, Meta plans to keep the LED light off during non-active saving scenarios such as "super sensing" or AI scanning, and only illuminate it when the user explicitly saves the audio or video. This setting may further intensify public concerns about continuous, unnoticed monitoring.
As a key point in the competition among tech giants for an AI hardware ecosystem, the advancement of round-the-clock wearable devices marks the transition of edge-side AI from "passive response" to "active perception." Meta's aggressive attempt not only reflects the intense conflict between AI hardware experience upgrades and traditional privacy boundaries, but also forces the industry to re-examine the balance between personal privacy protection and the widespread availability of AI services.
