According to a recent Swedish study, artificial intelligence could prove to be an invaluable tool during future global pandemics. This work suggests that AI could contain the increase in infections. The current context around Covid-19 has obviously stimulated this research, the theme of which had been explored very little in the past.
Today, the toll of the Covid-19 pandemic is more than three million deaths , including more than 550,000 in the United States alone. However, it will have been necessary to wait for the arrival of this pandemic to intensify research on a very little-known subject. In its study relayed in a press release of April 13, 2021, the University of Gothenburg (Sweden) indeed mentioned the beneficial uses of AI with the aim of quickly controlling an epidemic.
As part of this work, scientists evaluated how machine learning algorithms could develop effective screening methods. This would make it possible to better control future health crises, despite a limitation of data synonymous with a lack of visibility.
According to Laura Natali, a researcher in the Department of Physics at UG and one of the authors of the study, “this can be a first step towards better control by society of future major epidemics and reduce the need for restrictions and cessation of activities ”. Thus, there would not necessarily be a need to confine the population and weaken the economy by slowing down or stopping activities as is currently the case.
According to Swedish researchers, the new testing strategy based on machine learning could automatically adapt to the specific characteristics of sick people. Thus, the AI will define screening priorities according to the age groups of the population or certain geographical areas at different scales.
“When an outbreak begins, it is important to quickly and effectively identify infectious individuals. In random testing there is a significant risk of not achieving this, but with a testing strategy more focused on the goals one wishes to define, we can find more infected individuals and thus also obtain the information needed to reduce the spread of infection. We show that machine learning can be used to develop this type of testing strategy,” said Laura Natali.
This kind of solution could indeed see the light of day during a next large-scale epidemic . Recall that recently, US researchers mentioned a "spillover potential" in 500,000 pathogens while to date, we count "only" 250 viruses transmitted between animals and humans . In other words, there is no doubt that other pandemics of great importance will emerge. Everything will then depend on our ability to contain them.
