Amidst the ravage of COVID-19 world over, each week, hundreds of research papers published that report research findings. And, many findings reported do not undergo thorough peer review to evaluate their reliability.
So much so, in some cases, some research findings that have been poorly evaluated has massively influenced public policy. The reporting of curing COVID patients with a combination of azithromycin and hydroxychloroquine by a French team is an example of poor validation of research findings. So much was the claim publicized, soon patients in the U.S. were prescribed these drugs under authorization for emergency use. However, further research involving large number of patients casts serious doubts on such claims.
How can clinicians, researchers, and policymakers keep up with so much COVID-related information being circulated each week?
Meanwhile, in such a scenario, a team of scientists who work in AI companies opine that AI and machine learning have the potential to separate the wheat from the chaff.
Urgency for Coronavirus Vaccine resulting in sideline of Traditional Peer Review
In the current situation, the sense of urgency to develop a vaccine and develop effective treatments for the coronavirus has changed the way of working. Many scientists are bypassing the traditional peer review process by publishing preliminary versions of their work online.
“While this has enabled rapid dissemination of information, the problem arises when claims of drugs that have not been experimentally validated are published in preprint,” said a scientist fro University of New Mexico who opined of the potential of AI for COVID. Among other things, bad information may lead clinicians and scientists to waste time and money chasing worthless leads.
“AI and machine learning can utilize massive computing power to verify many of the claims that are being made in a research paper,” state a group of private and public sector researchers from few countries.