Synthetic Sniffer
Koniku is fusing biological and artificial neurons to sniff out cancer, COVID-19, explosives, and more.
With the rapid advancement of computing technologies and all the wonders it has brought us, it is easy to forget that our best efforts still pale in comparison to what biological machines are capable of. We previously reported on a method to hack the exquisitely sensitive olfactory system of the humble locust. Now, the company Koniku has developed a commercial product that fuses biology and silicon to sniff out signatures of trace elements that can signal the presence of explosives, COVID-19, cancer, and much more.
Koniku’s fried egg-shaped device, the Konikore, is powered by a processing core containing 128 active neurons. Initial applications of this core are for detecting scents, and Koniku claims it performs better than a dog’s nose. When the neurons detect a scent, the signals are processed with the help of machine learning (neurons of the artificial sort, that is). If the scent is determined to be one that the device was programmed to detect, it will light up.
While the device cannot directly sniff out COVID-19 or cancer, cells affected by such conditions release characteristic organic compounds that can be used to detect their presence. This is similar to the way in which dogs are able to detect a number of diseases. Training a dog for the task typically costs around $200,000, however the Konikore prototype is selling for around $3,000.
Koniku envisions their device as a low-cost, accessible means to diagnose disease in individuals or detect the presence of pathogens in public spaces. They also see applications in sniffing out explosives for airport security.
There are many questions around how this device will keep neurons alive and functioning normally for any long duration within this device. In vitro, neurons are fragile, and maintaining the correct balance of chemicals outside of their normal environment is a challenge. It is not clear how Koniku is approaching these problems, or if the device will function well in the wild. Perhaps these questions will be answered soon — the first deliveries are scheduled for December 2020.