Study shows standard WiFi signals can identify individuals with high accuracy, raising privacy concerns

Researchers from the Karlsruhe Institute of Technology (KIT) in Germany have demonstrated a new surveillance capability showing that ordinary WiFi networks may be used to identify individuals with extremely high accuracy. By analyzing how radio waves interact with people and objects in an environment, artificial intelligence systems can effectively reconstruct identifiable patterns of human movement and presence. Unlike earlier experimental approaches that required specialized equipment, this method relies on standard WiFi hardware already found in homes, businesses, and public spaces.

The technique makes use of beamforming feedback information (BFI), which is routinely exchanged between devices and routers during normal wireless communication. This data, often transmitted without encryption, can be intercepted and processed to generate multiple signal “views” of a person. Machine learning models trained on these signal reflections were reportedly able to recognize individuals with nearly 100% accuracy in tests involving 197 participants, regardless of their walking direction or viewing angle. Importantly, the system can identify people even if they are not carrying a device, since surrounding network activity still produces usable signal reflections.

Researchers warn that this capability could turn ordinary WiFi routers into passive surveillance tools, potentially enabling tracking in places such as cafes, offices, airports, or public streets without users being aware. They emphasize that while more invasive surveillance technologies already exist, WiFi is especially concerning due to its ubiquity and invisibility. The team highlights significant privacy risks and suggests that stronger safeguards should be incorporated into future WiFi standards, including IEEE 802.11bf. The findings underscore growing concerns about how everyday wireless infrastructure could be repurposed for large-scale, hard-to-detect monitoring of individuals.