How can active listening AI help improve public health?

Answer: By detecting coughs.

by / September 17, 2018

Sickweather builds real-time illness maps based on crowdsourced data. Mycroft AI’s Precise is an open source platform for a listening AI that can detect a single wake word in a public space. Together, they just might be the perfect team to address a request from the Kansas City, Mo., Health Department to develop a better way to track sickness in public locations.

Using the Precise framework, Mycroft and Sickweather trained an AI to single out coughs among other noises. Its accuracy was spot-on, and Kansas City was pleased with the initial demonstration. While it still has a long way to go to get to widespread deployment, the city sees promise in the technology.

“It reminds me of the ShotSpotter technology for detecting gunshots, and the implications for public health are just as exciting,” said Dr. Sarah Martin of the Kansas City, Mo., Health Department. “I imagine that it could not only provide the earliest warning for respiratory outbreaks but could eventually understand the nuance and patterns in types of coughs and their potential causes.”