Particle.news
Download on the App Store

AI Can Re-Identify People From Ordinary Wi‑Fi, KIT Researchers Warn

Unencrypted beamforming feedback turns routine router traffic into person-specific radio signatures.

Overview

  • In controlled tests with about 197 participants, the system re-identified individuals with near‑perfect accuracy after an initial training pass.
  • The method reads standard Beamforming Feedback Information exchanged by routers and devices, which is currently sent unencrypted on consumer hardware.
  • People can be recognized even if they carry no device, and any active WLAN node in range could act as a silent observer.
  • The KIT team urges standards bodies to add protections such as encrypting or restricting access to beamforming feedback, with proposals referenced for upcoming Wi‑Fi generations.
  • The findings, presented at the ACM Conference on Computer and Communications Security, include warnings that surveillance firms and intelligence services may pursue the technique because it is invisible and works through walls.