• 2 min read
Meta patent tracks mood through your voice
Meta patented a system that records voice, tracks mood and correlates emotions with location, activity and medication timing.
Image: TNW
Meta has patented an AI system that continuously records a user’s voice, transcribes it and uses a machine-learning model to estimate their emotional state. The patent was published on July 2 and was first spotted by Patentlyze.
The system listens for “audible communications,” including sighs, laughter, vocal tone, pauses and breathing. It combines those signals with contextual data such as time of day, location, physical activity and medication timing, creating a persistent mood log from “multimodal sensor inputs on synchronized timelines.”
Meta presents the technology as a fitness tool. According to the patent, emotional coaching could correct workout posture and adapt guidance in ways a human personal trainer cannot. But its design extends beyond exercise, describing “continuous emotional monitoring on everyday devices.”
The system could correlate mood with medication schedules and produce summaries such as “a happier emotional state associated with a particular time of day or at a time when medication is taken.”
Recommended reading
San Francisco orders Apple, Google to remove nudify apps
Amazon pursued a similar idea with its Halo Band, launched in 2020 with a microphone that performed “tone of voice analysis.” Following public backlash, Amazon removed the microphones from the next-generation model in 2021 and discontinued the entire product line in 2023.
Meta’s smart glasses are already facing privacy scrutiny over covert recording. The company has sold seven million pairs, while two US lawsuits allege that Meta misled consumers about how captured footage was handled. Persistent ambient voice recording would broaden that concern from what users see to how they feel.
A Meta spokesperson told 404 Media that:
“Patents at Meta are often filed to disclose concepts that may or may not be implemented.”
The patent’s detailed references to medication tracking and emotion-time correlations nevertheless describe a system far more specific than a basic workout feature. If it becomes a product, emotional recordings would also need to be labelled for model training—potentially requiring people to classify strangers' sighs, laughter and private speech by mood.
Security Editor
Sophia unpacks the invisible wars happening on our networks. Covering cybersecurity, privacy legislation, and cryptography, she exposes how our data is weaponized and defended. Before joining for(geeks), she spent years as a penetration tester. She's the reason the rest of the team uses physical security keys.
via TNW
