Researchers at the University of RomedevelopedA method purportedly can re-identify individuals based solely on the way their bodies interfere with Wi-Fi signals — a breakthrough that could reignite debates about privacy and surveillance. The technology is more powerful and less intrusive than biometric systems that rely on faces, fingerprints or mobile devices; and unlike fixed-location cameras or scanners, it can track individuals in any space covered by a Wi-Fi network.
The study's authors, computer scientists Danilo Avola, Daniele Pannone, Dario Montagnini and Emad Emam, describe WhoFi as a new type of biometric identification technology. It does not rely on visuals, wearables, or behavioral cues, but instead derives an individual's unique "signature" from changes in Wi-Fi channel state information, which captures changes in signal amplitude and phase as electromagnetic waves interact with physical obstacles.
"The core insight is that as a Wi-Fi signal propagates through the environment, its waveform is modified by the presence and physical characteristics of objects and people along the way," the authors wrote. "These changes...are rich in biometric information."
To test their hypothesis, the researchers trained a deep neural network to identify signal changes unique to individuals. The system learns to recognize different people by analyzing how each person modifies Wi-Fi signals, even in different environments. When tested on the NTU-Fi dataset, a widely used Wi-Fi-based human perception benchmark, WhoFi achieved up to 95.5% re-identification accuracy using a Transformer-based deep learning model.
This technology of using Wi-Fi for human body sensing is not new. Over the past decade, researchers have explored applications ranging from fall detection to sensing through walls and gesture recognition. EyeFi, a similar method proposed in 2020, reported that its person re-identification accuracy was about 75%. The authors of WhoFi believe that their method is more accurate and works reliably in different locations.
Although promising from a technical perspective, Wi-Fi-based re-identification technology raises serious ethical concerns. Unlike visible and identifiable cameras or RFID tags, Wi-Fi signals are ubiquitous and often difficult to detect because they are designed to transmit data rather than track. The Register noted that proponents of this emerging field argue that Wi-Fi sensing technology is a more privacy-focused alternative to visual surveillance because it does not capture images. However, critics argue that continued tracking - especially without the knowledge or consent of the person being monitored - could open the door to new covert forms of surveillance.
The team acknowledges the tension between innovation and privacy. While they insist that WhoFi does not directly capture personal identities or personal data, they are aware that it could be abused if deployed without proper protections.
Currently, this research is still in the academic stage and there are no plans for commercial or government applications. However, as Wi-Fi-equipped environments become more ubiquitous, the possibility of our bodies being able to silently transmit identity information without any equipment may soon move from experiment to reality.