Who-Fi is an innovative technology that utilizes artificial intelligence (AI) to recognize and monitor individuals without needing visual input. This experimental technology remains largely untested in real-world conditions. However, a recently released research paper outlines its proof of concept, suggesting its potential to convert standard Wi-Fi signals into a biometric scanning tool capable of tracking individuals’ movements and identifying their unique biometric signatures.
Understanding the Who-Fi Technology
A research paper available on arXiv details how conventional 2.4GHz Wi-Fi signals can be harnessed for individual identification and tracking, playing a crucial role in both identity verification and surveillance. This emerging technology raises significant concerns regarding digital privacy and security.
The Who-Fi system operates by combining Wi-Fi signals with a transformer-based neural network, also referred to as a large language model (LLM). This LLM interprets a phenomenon known as “channel state information” (CSI), monitoring fluctuations in Wi-Fi signal strength and phase as they interact with individuals within a space. This process is akin to how radar and sonar systems operate.
Whenever a person is in proximity to a Wi-Fi signal, the resultant alteration in the signal’s trajectory creates a distinctive pattern. This pattern purportedly boasts accuracy comparable to traditional biometric markers, such as fingerprints, facial features, and retinal structures. The Who-Fi system is designed to recognize these unique signatures and associate them with specific individuals.
After training on these biometric signatures, the system can track an individual’s movements and identify them even after a significant absence from the network zone. It also has the capability to collect data relating to body movements and recognize sign language. A key benefit of the Who-Fi system is its operation without the need for visual or auditory sensors, such as cameras and microphones.
The study reveals that the Who-Fi system comprises a single-antenna transmitter and a three-antenna receiver, allowing for cost-effective deployment. When it comes to functionality, researchers reported a remarkable precision rate of 95.5 percent, even when the target was located behind a wall and walking at a typical pace.
The system’s accuracy reportedly remains unaffected even if an individual changes clothing or carries a backpack. Notably, a single Who-Fi instance can identify and track up to nine people simultaneously.
Furthermore, Who-Fi demonstrates a high capacity for evasion, making it challenging for conventional surveillance technologies to detect it. This capability is facilitated by the lack of specialized hardware and the absence of distinctive emissions, infrared, radar, or visible light patterns that could signal its presence. Additionally, Who-Fi employs passive radio frequency (RF) sensing, enhancing its ability to operate discreetly.
