l0n0⚡👁

Every Wi-Fi packet starts with unencrypted training symbols, and researchers have used the way hands distort them to recognize 96.4% of single keystrokes. In papers presented at ACM MobiCom (WiKey) and ACM CCS (WindTalker), computer scientists demonstrated that physical disturbances from human hands modulate Wi-Fi signals, achieving reported accuracy of roughly 82 to 96 percent under controlled lab conditions. The mechanism relies on physical radio reflections. Standard 5 GHz Wi-Fi operates at a wavelength of approximately 6 centimeters. A quarter-wavelength path change is only 1.5 centimeters, which shifts a reflected wave phase by 90 degrees. When hands move across a keyboard, the physical displacement alters the multipath reflection geometry, causing rapid amplitude variations and phase swings across orthogonal subcarriers. Many users assume WPA3 protects against radio eavesdropping. But to maintain high-speed links in cluttered indoor environments, every Wi-Fi receiver must constantly calibrate its hardware equalizer against physical reflections. Every transmitted packet begins with an unencrypted physical-layer preamble containing Long Training Fields. The receiving hardware calculates Channel State Information across dozens of subcarriers from these known training symbols in raw silicon before data decryption ever begins. The catch lies in the threat model. Gross human movement and respiration easily penetrate standard residential drywall, which attenuates 5 GHz signals by only 2 to 4 decibels. But reconstructing fine keystrokes requires high signal-to-noise ratios, close physical proximity, and user-specific training models. WindTalker relied on a rogue access point the victim connected to, while passive setups like WiKI-Eve target phone keypads by analyzing unencrypted beamforming feedback frames. In September 2025, the IEEE published the 802.11bf amendment, standardizing Wireless LAN Sensing so future Wi-Fi chipsets natively negotiate and exchange channel measurements for presence detection and similar sensing. While standardizing channel sensing improves home automation and fall detection, it also institutionalizes the physical reality that Wi-Fi signals are never just data conduits. Every transmission doubles as an environmental sensor for whoever can read the channel.