Meta is testing a prototype of super sensing AI glasses that would continuously record audio and capture photos every few seconds, letting wearers query their day through AI. The $1.5tn company's executives are planning not to activate the LED indicator that currently signals recording on its Ray-Ban smart glasses, making it harder for bystanders to know they're being filmed. Raw footage wouldn't be stored; only metadata would upload to Meta's servers for AI processing, though the company is also weighing whether to use collected data to train its own models in the race against OpenAI and Google.

The push reflects CEO Mark Zuckerberg's vision of AI glasses replacing smartphones as the primary AI interface. Meta acquired Limitless, maker of AI pendants that transcribe conversations, in December, doubling down on always-on hardware after abandoning its costly metaverse headset strategy. Cheaper Ray-Ban variants, including frames designed by influencer Kylie Jenner, have driven recent adoption, while advanced models now overlay displays on lenses.

Legal exposure is substantial. Wiretapping laws in multiple U.S. states prohibit recording audio without consent, and biometric data regulations could apply. Boston University law professor Woodrow Hartzog warns that no single law addresses always-on devices. Meta's existing glasses already face scrutiny: contractors in Kenya viewed graphic user content, and unreleased facial recognition code was discovered in the platform last month.