The dangers are obvious (and also there are some fascinating insights into how healthcare works practically in China). I wonder if some kind of "second opinion" antagonistic approach might reduce the risks.
Can you expand on the insights into the chinese medical system ?
Medical Advice Generative Adversarial Networks would be a good idea.
I see some of this adversarial second-guessing introspection from Claude sometimes. ("But wait. I just said x y and z, but that's inconsistent with this other thing. Let me rethink that.")
Sometimes when I get the sense that an LLM is too sycophantic, I'll instruct it to steelman the counter-argument, then assess the persuasiveness of that counter-argument. It helps.