Early within the pandemic, quite a lot of researchers, startups and institutions developed AI systems that they claimed could diagnose COVID-19 from the sound of an individual’s cough. On the time, we ourselves were keen about the prospect of AI that may very well be yielded as a weapon against the virus; in a single headline, we endorsed cough-scrutinizing AI as “promising.”

But a recent study (first reported on by The Register) suggests that some cough-analyzing algorithms are less accurate than we — and the general public — were led to consider. It serves as a cautionary tale for machine learning tech in healthcare, whose flaws aren’t at all times immediately apparent.

Researchers from The Alan Turing Institute and Royal Statistical Society, commissioned by the U.K. Health Security Agency, conducted an independent review of audio-based AI tech as a COVID-19 screening tool. Along with members from the University of Oxford, King’s College London, Imperial College London and University College London, they found that even probably the most accurate cough-detecting model performed worse than a model based on user-reported systems and demographic data, resembling age and gender.

“The implications are that the AI models utilized by many apps add little or no value over and above the predictive accuracy offered by user-reported symptoms,” the co-authors of the report told TechCrunch in an email interview.

For the study, the researchers examined data from greater than 67,000 people recruited through the National Health Service’s Test and Trace and REACT-1 programs, which asked participants to send back nose and throat swab test results for COVID-19 together with recordings of them coughing, respiratory and talking. Using the audio recordings and test results, the researchers trained an AI model, attempting to see whether coughs could function an accurate biomarker.

Ultimately, they found that they might not. The AI model’s diagnostic accuracy wasn’t significantly better than probability when controlling for confounders.

Partly responsible was recruitment bias within the Test and Trace system, which required participants to have no less than one COVID-19 symptom as a way to participate. But professor Chris Holmes, lead creator of the study and program director for health and medical science at The Alan Turing Institute, says the findings show coughs are a poor predictor of COVID-19 normally.

“It’s disappointing that this technology doesn’t work for COVID-19,” he told TechCrunch in an emailed statement. “Finding recent ways to quickly and simply diagnose viruses like COVID-19 is actually essential to stop its spread.”

The study is a blow to business efforts like Fujitsu’s Cough in a Box, an app funded by the U.K.’s Department of Health and Social Care to gather and analyze audio recordings of COVID-19 symptoms. And it puts some scientific claims unsure. One paper co-authored by researchers on the Massachusetts Institute of Technology pegged the accuracy of a cough-analyzing COVID-19 algorithm at 98.5% — a percentage that looking back seems dubiously high.

That isn’t to suggest the Turing Institute study is the last word on cough detection where it concerns COVID-19. Holmes leaves open the likelihood that the tech may match for other respiratory viruses in the long run.

However it wouldn’t be the primary time healthcare AI has overpromised and underdelivered.

In 2018, STAT reported that IBM’s Watson supercomputer spit out erroneous cancer treatment advice, the results of training on a small variety of synthetic cases. In a newer example, a 2021 audit of healthcare system provider Epic’s AI algorithm for identifying patients with sepsis was found to miss nearly 70% of cases.