Two articles, one from a radiology perspective, one from a more general perspective, highlighting the need to proactively consider governance of use of analytics/ML/AI in medicine. In particular, the highlight the current deficits of regulatory agencies in oversighting these products prior to approvals. Apologies for the scribbles on one - likely removable by someone more savy than me.
I think that was some of the concern particularly around the Singapore hack. There have been a number of reports of theoretical use of data injection to screw up algorithms in healthcare data. Similar to the insertion of white lines randomly on the road causing the self-driving cars to drive off the road or into opposing lanes.