A recent study has revealed a concerning new surveillance threat: even without using Wi Fi or even turning off your phone, attackers may still be able to "detect people through the walls" of wireless networks around you. A team from the KASTEL Institute for Information Security and Reliability at Karlsruhe Institute of Technology in Germany has warned that attackers can use existing Wi Fi signals in the environment to accurately infer the identities of individuals present by passively monitoring changes in radio wave propagation, without the need for any special hardware. The research paper was published in the latest issue of ACM Data Science journal. The team members explained that by observing how Wi Fi radio wave signals are reflected and disturbed by the human body, one can "see" the surrounding environment and people inside like a camera imaging. Therefore, whether people carry or turn on Wi Fi devices is not important. Even if the device is in off or airplane mode, as long as it is in a space with Wi Fi signals, it may become a recognized target. The key to this technology is that devices in modern Wi Fi networks regularly send a signal called beamforming feedback information to routers. These signals were originally used to optimize network connection quality to ensure more accurate transmission to user devices. But these feedback data are not encrypted during transmission, and anyone within the signal range, including malicious listeners, can intercept them. By analyzing how these signals undergo subtle changes due to human movement and posture, combined with machine learning models, the motion characteristics of the human body can be constructed, and specific individuals can be identified. This means that every Wi Fi router may become an observer. For example, when people pass by a coffee shop that provides Wi Fi every day, even if they have never connected to its network, their walking posture and physical characteristics may still be recorded and used for subsequent recognition. This type of monitoring can be utilized by cybercriminals to achieve long-term tracking. Unlike previous monitoring technologies that relied on LiDAR or specialized sensors, this new method does not require expensive or very obvious equipment. Attackers only need a regular Wi Fi receiver to use existing legitimate communications in the environment for monitoring. Moreover, this danger lies in its "invisibility", which is not only difficult to detect, but also not limited by physical line of sight. In an experiment involving 197 participants, the team achieved nearly 100% accuracy in identity recognition through only a few seconds of signal analysis, and was not affected by observation angle or personal walking style. This high-precision recognition capability not only highlights the powerful technological potential, but also reveals the serious privacy risks involved. (New Society)