The Financial Times (FT) reported on Wednesday that Meta is testing a prototype of super sensing smart glasses, which keep the camera and audio on continuously throughout the day.
Sources reveal that Meta is developing smart glasses capable of capturing photos every few seconds while recording ambient sounds. Users can leverage artificial intelligence (AI) to inquire about or recall their daily experiences.
The controversy stems from Meta executives’ plans to disable the light-emitting diode (LED) indicator when the super sensing feature is active.
Current AI smart glasses typically illuminate an LED on the frame to signal when the wearer is taking photos or recording video. Disabling this indicator would make it challenging for others to detect if they’re being recorded, raising significant privacy concerns.
Even with existing smart glasses that have active LED indicators, some users have sparked controversy by using online-purchased LED-blocking stickers to conceal the light, effectively turning the devices into covert cameras.
The FT reports that this has ignited an internal debate at Meta over privacy protection measures.
Meta had previously gone all-in on its metaverse vision, even rebranding the company, but after substantial losses, it pivoted to focus on AI and wearable device development.
Currently, Meta is ramping up its efforts in the AI wearable market, centered on its Meta Ray-Ban smart glasses series, developed in partnership with eyewear giant EssilorLuxottica.
During Meta’s first-quarter earnings call, Chief Executive Officer (CEO) Mark Zuckerberg expressed his vision for the company’s smart glasses to evolve beyond just answering questions, becoming an all-day companion that remembers experiences and assists users in achieving their goals.
As Meta vies with other tech giants in the AI arms race, there’s speculation about whether it might leverage data collected from smart glasses wearers to train its proprietary AI models.
Experts caution that Meta’s always-on recording device in development likely violates data protection and privacy laws. Similar to South Korea, many U.S. states prohibit recording someone’s voice without consent.
Woodrow Hartzog, a professor at Boston University School of Law, told the FT that there’s no comprehensive legislation addressing how these technologies are designed and implemented. Lawmakers need to take this seriously and update the legal framework to address these new realities.
Inside sources suggest that Meta is considering an alternative approach: not providing users direct access to images and audio captured through the super sensing feature. This data would not be stored on Meta’s servers either.
Instead, when metadata from images and audio is uploaded to the server, Meta’s AI could process it to provide services to users. Some Meta developers argue this approach could mitigate privacy concerns.