Think your private phone call is safe because it’s encrypted? Think again. Researchers at Penn State have demonstrated that radar—a technology more commonly used for tracking planes and detecting cars in your blind spot—can be repurposed to eavesdrop on smartphone conversations by detecting microscopic vibrations emanating from the device itself.
The system, dubbed WirelessTap, utilizes commercially available millimeter-wave radar to capture the barely perceptible tremors that a phone’s speaker creates while in use. These vibrations, just 7 micrometers in size, are invisible to the human eye but visible to radar. Once detected, they can be reconstructed into sound and partially transcribed with the help of artificial intelligence.
In controlled lab tests, WirelessTap achieved nearly 60% word accuracy at 20 inches, while accuracy at 10 feet dropped to 2–4%. That might not sound like much, but even fragments of conversations can reveal credit card numbers, passwords, or key phrases when paired with context. When phones were held by users rather than mounted, accuracy dipped but still reached around 40% at short range—enough to piece together meaningful snippets.
The researchers stress that this isn’t about hacking data networks or breaking encryption. Instead, it’s like watching ripples in a glass of water and inferring what was said nearby. AI speech recognition systems, like Whisper, make filling in the gaps easier than ever.
The risks? Corporate espionage, identity theft, and privacy breaches. The solutions? Potential defenses range from adding vibration-dampening materials around earpieces to building in “noise” signals that mask real vibrations.
For now, the threat to the average person is low—WirelessTap requires specialized equipment and a direct line of sight. However, as radar sensors become cheaper and AI becomes sharper, the study serves as a clear warning: privacy isn’t just about data, but also about the physical signals our devices leak every day.