WiFi can be used as a biometric because the human body subtly changes radio signals as we move. These changes, captured as Channel State Information (CSI) by ordinary WiFi devices, contain distinctive patterns related to how people walk, move, and occupy space.

Research shows that an individual’s gait can be identified with high accuracy using machine‑learning models, providing a contactless alternative to cameras and wearables. In this talk, we explore what WiFi can and cannot do for identification and recognition, and discuss the associated limitations and privacy concerns.