This technology, named “WirelessTap”, uses millimeter wave sensors (mmWave) to capture the tiny vibrations produced by a phone during calls and convert them into audio that can be transcribed, according to a report by “New Atlas”. The technology relies on radio waves in the 77-81 GHz range, commonly used in applications such as motion detection within homes or assisting autonomous vehicles.
Suryoday Basak, the lead researcher in the study published in the Proceedings of WiSec journal, explained that using machine learning algorithms, such as the Whisper model from OpenAI, allows these vibrations to be converted into understandable speech. He added, “By incorporating the context of the conversation, accuracy can be improved, just as lip readers can understand speech more accurately when they know the topic.”
Although the technology has not yet reached perfection, it raises serious privacy concerns. WirelessTap surpasses previous radar eavesdropping technologies that relied on speakers or were limited to specific vocabularies. Basak warned that this technology could be used by malicious entities to steal private information, emphasizing that the study’s goal is to raise public awareness of the risks of remote eavesdropping.
He affirmed, “We hope these findings will encourage individuals to be cautious during sensitive calls.” This study highlights the dark side of technological advancement, where innovations can be used in ways that threaten privacy.
