AI wearables reignite privacy concerns: Amazon’s Bee and Meta’s smart glasses’ “always-listening” feature raise legal and trust issues. According to Bloomberg, a bug has been identified in Meta Ray-Ban smart glasses. Multiple users reported on Reddit that the glasses continued to record audio even when the “Hey Meta” wake word was not spoken and the “voice activation without wake word” setting was disabled. Audio recordings and their text transcriptions were saved to the Meta AI app, and some users observed the glasses’ white indicator light flickering intermittently. Meta acknowledged the issue and stated it had been “fixed,” but did not provide further details. Users can now view, manage, or delete these voice recordings in their settings. Meanwhile, Amazon has launched its Bee series of always-listening devices, designed to continuously monitor conversations, transcribe them, and summarize them into to-do lists and reminders to help users “review their day.” Amazon claims that audio is deleted after processing, not stored long-term, and is encrypted with access restricted solely to the user. However, privacy experts note that even without retaining raw audio, continuously receiving and interpreting spoken conversations may still violate wiretap laws in certain U.S. states that require two-party consent—potentially exposing the wearer to legal liability, as bystanders have not given consent. This year, Apple, Meta, Amazon, and several startups have all introduced AI devices with “always-listen/understand” capabilities. These products shift the privacy boundary from “record only after a wake word” to “constantly listen and summarize,” sparking ongoing debates over legal, ethical, and user trust issues. Meta’s glasses have previously drawn scrutiny over contractor access to private footage and class-action lawsuits.