Our society is increasingly becoming a surveillance state across the globe. The number of active cameras in the world that record everything we do in the public have been exponentially increasing every year and it is now possible to follow a person across locations and track them easily. Systems have been using gait analysis, facial recognition etc to identify folks and now they have a new way to identify (or re-identify) people using Wifi. Researchers(Danilo Avola, Daniele Pannone, Dario Montagnini, and Emad Emam, from La Sapienza University of Rome) in Italy have developed a way to create a biometric identifier for people based on the way the human body interferes with Wi-Fi signal propagation and claim to have reached a 95.5% accuracy.
In 2020, the Wifi Alliance approved the IEEE 802.11bf specification that supported Wi-Fi Sensing which used existing Wi-Fi signals to sense motion amongst other things and routers with this capability are available in the market already. This study expands the Wi-Fi Sensing capabilities by using the Channel State Information (CSI) of a Wifi signal to distinguish individuals based on how their bodies alter signal waveforms. By learning the patterns from CSI sequences, the study claims to perform Re-ID by capturing and matching these radio biometric signatures.
“The core insight is that as a Wi-Fi signal propagates through an environment, its waveform is altered by the presence and physical characteristics of objects and people along its path,” the authors state in their paper. “These alterations, captured in the form of Channel State Information (CSI), contain rich biometric information.” CSI in the context of Wi-Fi devices refers to information about the amplitude and phase of electromagnetic transmissions. These measurements, the researchers say, interact with the human body in a way that results in person-specific distortions. When processed by a deep neural network, the result is a unique data signature.
Researchers proposed a similar technique, dubbed EyeFi, in 2020, and asserted it was accurate about 75 percent of the time. The Rome-based researchers who proposed WhoFi claim their technique makes accurate matches on the public NTU-Fi dataset up to 95.5 percent of the time when the deep neural network uses the transformer encoding architecture. “The encouraging results achieved confirm the viability of Wi-Fi signals as a robust and privacy-preserving biometric modality, and position this study as a meaningful step forward in the development of signal-based Re-ID systems,” the authors say.
The study claims are impressive but I am skeptical about the claims in it, primarily because it is quite easy to modify how Wifi signals propagate through your body. For example, I can carry a metal mesh rolled up in my pocket and then later on open it up and put it around my ribcage. I have immediately modified how the WiFi signal passes through the body and the study doesn’t go into details on how it would work in that scenario or other similar cases. In fact spraying metal infused water on myself would also change how the signal interacts with my body.
They claim this is more privacy preserving because it doesn’t show the face or body but I feel it is worse because it allows (if it works) folks to track a person with good accuracy across locations. Which makes it a powerful surveillance tool. I can imagine it being deployed in restrooms of companies like ‘Three Brothers Machine Manufacturing’ in China who have strict bathroom break policies (two-min max) to ‘boost efficiency’, as it will allow them to monitor who is inside a bathroom without having active camera’s in the bathroom.
Facial recognition is already flaky in real world use with a high error rate of 34.7% for darker-skinned people, according to a 2018 study titled “Gender Shades” by Joy Buolamwini and Timnit Gebru. People have been arrested after being falsely identified by facial recognition systems and I feel that if this WhoFi system gets deployed in large scale we will see similar issues with it as well.
– Suramya
