Such information as this, tweeted by Gregory Conley, is important to keep within your awareness:

Such exploration is valuable, but we need strong personal and cultural safeguards against abuse.  Earlier this week, the Dall-E 3 AI (via ChatGPT) I use for many of the images on this site refused to add the Confederate flag on top of the General Lee from the Dukes of Hazzard.  Even putting aside errors and the sci-fi talk of AI’s developing its own hostile intentions, we have to wonder what medical AI will refuse to do or, conversely, insist on doing.

On the human level, talk about how “using a human [for medical treatment] *at all* is costing lives” is a plea to shift diagnosis away from a human being in whose eyes you can look to other human beings who may have no interpersonal medical experience at all.