COVID-19 has led to inundated doctors using inadequately tested artificial intelligence (AI) tools. While AI-enabled tools can detect the virus quicker than traditional tests, inadequately-tested tools could cause undue harm to patients, says GlobalData a leading data and analytics company. 

AI technologies are increasingly being used to predict and diagnose COVID-19 and hospitals are keen to adopt these new technologies thanks to their ability to deliver quick results. Tools being tried out include AI-enabled stethoscopes, computed tomography (CT) imaging that incorporates AI, and MIT and Harvard’s research that uses algorithms to diagnose COVID-19 using voice analysis. While high-speed COVID-19 AI CT imaging analyses might be desirable, a high level of proven clinical validity over time is the only real way to prove its ability to diagnose, and potentially control the spread of the virus.

Shabnam Pervez, Technology Analyst in Medical and Healthcare, Thematic Research, at GlobalData comments: “To keep up with demand for COVID-19 testing, new tools are being rushed into service and are not undergoing the kind of rigorous testing they would receive under normal circumstances.

“AI-enabled devices should be used with precaution and only as a guideline for clinical intervention for now. AI-enabled tools’ ability to test COVID-19 is still in the early stages of implementation and output should not be interpreted as a final concrete result, but rather as a guide for the next course of action.”