No camera, no consent, just WiFi watching you

Although it is a breakthrough in science, the method has sparked controversy about privacy and ethical surveillance.

Here’s how it works: researchers at La Sapienza University of Rome see the way a body disrupts a WiFi signal as a kind of invisible fingerprint – it’s unique to everyone. However, this is not the traditional biometrics we’re used to, such as facial recognition on our phones or fingerprint scanning to enter the gym. This method doesn’t require any cameras or even the person to be present.

Instead, the system relies on variations in WiFi signal patterns, specifically, in channel state information (CSI) that changes as electromagnetic waves bounce off and pass through the human body. These subtle disruptions are caused by the person's body shape, size, and movement. As well as a fingerprint, they create a unique "signature" in the signal’s amplitude and phase.

In order to decode these “signatures,” the team trained a transformer-based deep neural network on the NTU-Fi dataset, which is a standard in WiFi-based human sensing, says the report. The model learned to re-identify individuals with up to 95.5% accuracy, even as they moved through different spaces.

Unlike fixed-location cameras or biometric scanners, this system can operate anywhere a WiFi signal is present, possibly turning private homes, offices, or any other public space into a potential zone of passive identification.

While the system doesn’t store images or traditional personal data, its ability to track people without their consent or awareness has already raised major ethical questions, especially as WiFi networks become denser.