US develops artificial intelligence algorithm that can identify Covid-19
- The programme can diagnose people with coronavirus, even those without symptoms, based on recordings of their coughs
- The algorithm was built by the Massachusetts Institute of Technology but needs to be regulated before it is made into a free app
The US has developed an artificial intelligence (AI) algorithm that can correctly identify people with Covid-19 by using recordings of their coughs.
A report by the IEEE Journal of Engineering in Medicine and Biology says the AI filtered through thousands of recordings and overall achieved a 98.5 per cent success rate for people who tested positive for the virus. That number rose to 100 per cent for people who did not show any symptoms.
The algorithm was built by the Massachusetts Institute of Technology (MIT), in the US, but needs to be regulated before it can be developed into a free app.
Is it the cold, flu, or Covid-19?
Brian Subirana, MIT scientist and co-author of the report, told the BBC: “The way you produce sound changes when you have Covid, even if you’re asymptomatic.”
The difference between a normal cough and a cough from an asymptomatic patient could not be detected by the human ear.
“The effective implementation of this group diagnostic tool could diminish the spread of the pandemic if everyone uses it before going to a classroom, a factory, or a restaurant,” said Subirana.