Meta Removes Face-Recognition Code Following Privacy Scrutiny

Meta has removed unreleased face-recognition software from its Meta AI companion app just one day after a WIRED investigation revealed the code's presence. The system, internally referred to as "NameTag," was designed to convert images captured by Meta’s smart glasses into biometric faceprints, which would then be compared against a local database. The latest app update effectively purged the libraries, folders, and alert triggers associated with this functionality.

Prior to this removal, Meta had publicly characterized the feature as purely exploratory, with company executives labeling reports of its integration as misleading. However, the presence of the code in an app installed on over 50 million devices suggests that the infrastructure for biometric identification was significantly more developed than the company’s public statements implied. The system was reportedly capable of cropping, indexing, and storing images of individuals who could not be immediately identified, raising significant concerns regarding user privacy and the potential for misuse.

This incident highlights the growing tension between rapid technological development and the ethical implications of biometric surveillance. Privacy advocates have long warned that such tools could facilitate stalking or harassment by allowing users to identify strangers in public spaces without consent. By quietly embedding these capabilities into widely used software, Meta has reignited debates over transparency and the necessity of robust oversight for wearable AI devices.

While Meta maintains that no final decisions have been made regarding the deployment of NameTag, the swift removal of the code following public exposure underscores the sensitivity surrounding facial recognition technology. As Meta continues to integrate AI into its hardware ecosystem, the company faces increasing pressure to clarify its data retention policies and provide clear opt-in mechanisms for features that fundamentally alter the privacy expectations of the public.