Responsible AI in Medicine
Here are some resources and musings about how to do responsible AI in the field of medicine.
The following are some resources I collected during my participation in the Responsible AI in Medicine workshop (Copenhagen, 2022).
My own presentation in the workshop was a poster about teaching analytics medical data common sense. See here. The paper details common errors found in electronic HEalth Records and how to find them.
If you want to make sure you are reporting your findings in a non-biased and responsible manner, there are several frameworks to choose from. In medical research in general it is common to refer to the equator network for these things. Specifically, for ML and AI, the TRIPOD guidelines are relevant.
For (mostly bad, a few good) examples from recent COVID papers, see Laure Wynants' work here.
Some general best practices for ML research are nicely summarized here.