Meta Allows Opt-Out on Smart Glasses Visual Data Training

On September 24, 2026, Meta announced that users of its smart glasses can now opt out of having their "visual data" used for AI training. As reported by Engadget, this marks another concession to privacy advocates, promising that opted-out visual data will not be used for training or reviewed by third-party contractors.

Previously, while direct photos and videos captured by the glasses were not utilized for training, interactions with Meta AI that leveraged the camera were fair game. For example, asking "Hey Meta, what does this sign mean?" would capture and store the visual context. Engadget notes this distinction was often unclear to users, as these captures were not saved to a personal camera roll.

Further complicating matters, some of these visual recordings were reportedly sent to third-party contractors for review and labeling, a practice detailed in an earlier Engadget report. These contractors, often based in other countries, allegedly encountered "personal and intimate images" during their work, raising significant privacy red flags.

The newly introduced opt-out specifically addresses this visual data concern. However, it's crucial to note that audio interactions with Meta AI are still subject to training and review, though users can manually delete those recordings. This policy shift arrives amidst ongoing scrutiny of Meta's smart glasses privacy practices, including a proposed class-action lawsuit, and coincides with the reveal of new Ray-Ban branded glasses featuring no cameras at all.

Our take: Meta's decision to implement a visual data opt-out, especially after previous denials and obfuscations, underscores the immense pressure from privacy advocates and legal challenges. This is a reactive, not proactive, move. While a welcome improvement, the carve-out for audio data means users must remain vigilant about their interactions, as 'opting out' is rarely a comprehensive solution when companies retain granular control over data streams.