Are You Really Muted?: A Privacy Analysis of Mute Buttons in Video Conferencing Apps

Video conferencing apps (VCAs) make it possible for previously private spaces — bedrooms, living rooms, and kitchens into semi-public extensions of the office. For most part, users have accepted these in their personal space without much thought about permission models that govern use data during meetings. While access to a device’s video camera is carefully controlled, little has been done ensure same level privacy accessing microphone. In this work, we ask question: what happens microphone when user clicks mute button VCA? We first conduct study analyze users’ understanding model button. Then, using runtime binary analysis tools, trace raw audio flow many popular VCAs as traverses app from driver network. find fragmented policies dealing with among some continuously monitor input mute, others do so periodically. One transmits statistics its telemetry servers while muted. Using network traffic intercept en route server, implement proof-of-concept background activity classifier demonstrate feasibility inferring ongoing meeting cooking, cleaning, typing, etc. achieved 81.9% macro accuracy on identifying six common activities intercepted outgoing packets