Yep, exactly.

As a doctor who’s into tech, before we implemented something like AI-assisted diagnostics, we’d have to consider what the laziest/least educated/most tired/most rushed doctor would do. The tools would have to be very carefully implemented such that the doctor is using the tool to make good decisions, not harmful ones.

The last thing you want to do is have a doctor blindly approve an inappropriate order suggested by an AI without applying critical thinking and causing harm to a real person because the machine generated a factually incorrect output.