Meta smart glasses face privacy scrutiny over AI training and recording concerns

Meta's smart glasses are facing intense scrutiny from lawmakers and privacy advocates regarding the potential for nonconsensual recording and the use of intimate imagery to train artificial intelligence. US Senators Ron Wyden and Edward Markey have issued a letter alleging that Meta's visual AI features may retain imagery used for AI development, potentially exploiting sensitive information without bystander consent.

In response to privacy concerns in Hong Kong, Meta officials stated that the glasses include a built-in detection mechanism: if the recording indicator light is covered or physically tampered with, the camera will stop recording or disable itself. Meta also noted that its AI visual queries automatically blur faces and license plates to protect privacy.

However, investigations in China have revealed that inexpensive black stickers can be used to cover the indicator lights on various mainstream smart glasses, including models from Huawei and Xiaomi, allowing for covert recording without triggering alarms. Experts are calling for mandatory hardware standards to ensure that recording automatically ceases if the indicator light is obstructed.