While artificial intelligence has been a massive failure in helping the world to cope with the pandemic, it’s not all bad news.
We’ve known for quite some time in 2020 that processing vocal recordings of infected but asymptomatic people reveals potential indicators of Covid-19. This would be impossible without algorithms since the changes are too subtle for the human ear to notice.
While Bluedot gets some acolytes for early detection of Covid-19, it’s a simple Voice AI test where AI can truly shine with the pandemic. Dr. Kamran Khan of Bluedot used big data and artificial intelligence to warn the world of a potentially serious viral outbreak three days before the World Health Organization, though they picked up on the signs even earlier.
With AI in Covid-19 testing though we have to hand things over to MIT. MIT Lincoln Laboratory researchers are detecting these changes in Covid-19 patients even when these changes are too subtle for people to hear or even notice in themselves. Basically, researchers found that changes in the movements of vocal muscles could be an indicator of whether someone has suffered an inflammation from Covid-19.
All thanks to AI soon we’ll have more efficient and streamlined ways of testing. Testing for Covid-19 currently involves a deep throat saliva sample or the dreaded nasal swab. But it might soon be as simple as listening to the sound of your voice.
By some estimates, asymptomatic carriers of the new coronavirus make up 40 per cent of all cases. Researchers at MIT believe changes in people’s voices might indicate whether or not they have been infected.
These biomarkers stem from disruptions the infection causes in the movement of muscles across the respiratory, laryngeal, and articulatory systems. A technology letter describing this research was recently published in IEEE Open Journal of Engineering in Medicine and Biology.
Artificial intelligence can truly help us during the pandemic, we just have to be smart about it how it is used. Places like Taiwan, China and South Korea have used Big Data apps to facilitate contact tracing in a way North and South America have failed to do. This is also partly cultural due to privacy concerns.
Vocal biomarkers certainly do exist for Covid-19, since it can impact the brain and lungs and leave subtle symptoms. The research just needs to be fast-tracked and implemented into a widely adopted consumer app.
To conduct the test, the researchers used videos of five carriers from press conferences and interviews posted to YouTube, Instagram and Twitter. They analysed audio data from both before and after the carriers tested positive for the virus.
By using algorithms to analyse vocal signals, researchers detected disruptions in a person’s voice that could be caused by changes in movements of the larynx and muscles in the respiratory system. AI hasn’t been a deal breaker in helping us during the pandemic, but that’s not to say it couldn’t be.