The rapid advancement of AI is creating new technological possibilities, but also intensifying privacy concerns. Who-Fi is a new technology that demonstrates this, using AI to identify and track individuals without the need for cameras. This approach analyzes Wi-Fi signals to monitor activities and recognize people, raising important questions about digital security.
According to research published in arXiv, Who-Fi uses standard 2.4 GHz Wi-Fi signals to identify and track individuals. The system combines Wi-Fi signals with transformer-based neural networks. The neural networks analyze changes in the Wi-Fi signal, known as channel state information, to see how it is affected by a person’s presence, in a process that resembles radar or sonar.
When a person interacts with a Wi-Fi signal, it creates a unique pattern, which can be used for identification. After training, the system can track activities, re-identify individuals, and even interpret sign language. The key feature of this system is that it works without cameras or microphones. The study notes that it needs a single-antenna transmitter and a three-antenna receiver, making it relatively easy and inexpensive to implement.