Meta acknowledged a vulnerability in its Ray-Ban Meta smart glasses, which allowed the recording of conversations even without issuing the activation command 'Hey Meta'. The captured audio was subsequently made available in the Meta AI application, along with its respective textual transcriptions.
These reports began circulating on the Reddit platform following a recent firmware update to the glasses, as reported by TechSpot. Although the company did not detail the cause of this flaw, it assured that the issue has been resolved.
The company stated: 'This issue has been resolved. People can manage, review, and delete voice recordings from their interactions with Meta AI through the settings at any time.'
The Ray-Ban Meta models are available for sale in Brazil starting at R$ 3,699.
One user reported finding recorded conversations even when the 'Respond without Hey Meta' feature, which allows certain interactions to continue without repeating the activation phrase, was disabled. This individual mentioned that the white indicator light illuminated during some of these recordings, making this LED a crucial point in the company's security mechanisms for signaling the operation of capture features.
The owner claimed to have received guidance from Meta support, suggesting the deactivation of camera and microphone permissions, detection of glasses usage, deletion of command history, and temporary suspension of automatic updates. However, he observed that such actions could impact essential device functionalities and informed that he is consulting lawyers about the possibility of filing a class action lawsuit against Meta.
This incident adds to other privacy concerns related to Meta's camera and microphone-equipped glasses. There have been cases where users exploited the camera's discretion to film other people on social media without proper consent.
Additionally, information from third-party workers responsible for analyzing the AI-generated recordings from the devices indicated that the camera could be capturing consumers in private moments, including scenes of a sexual nature. According to TechSpot, this material was reviewed during the training of Meta's artificial intelligence system.
The controversies have forced the company to constantly improve its security systems, particularly regarding the LED that signals an ongoing recording. The glasses now prevent recording when the system detects that the light is obstructed or damaged. In September, the company informed Tecnoblog that it had deactivated 0.1% of sold glasses that presented tampering.
Despite this, the glasses continue to be the subject of global debates on regulations and legislation, especially concerning their introduction in certain environments. In Brazil, the Superior Electoral Court (TSE) has already listed them among the items prohibited in voting booths, while the Consumer Defense Institute (Idec) filed a complaint with the Federal Public Ministry to initiate an investigation into the risks of these devices.