Use of machine learning for COVID-19 diagnosis

Recently I came across a very interesting article that talked about the performance of 300 odd Machine Learning models that were developed in 2020 for COVID-19 diagnosis. Researchers concluded that none of the models were suitable for this purpose. This was owing to multitudes of reasons such as technological (quality of images), process (biases, methodological flaws), data (insufficient data) etc. Reading through it, got the impression that having a good amount of representative prior data remains an inescapable pre-requisite for successful Machine Learning models.

Given below is the link to that article. Have a look and let me know if you infer something else from it.

Machine learning for Covid-19 diagnosis: promising, but still too flawed (healthcare-in-europe.com)