Comment by 2muchcoffeeman

I think his last point is important

* Could the AI solution be somehow misaligned by exploiting a weakness in the trial process or its math models?*

Maybe it’s a clumsy example but I think the difference is humans would follow the spirit of the trial process.

Could an AI find flaws in the process and then use them to get its drug passed? I think that’s the question.

Are you saying that the trail process is good enough that we shouldn’t worry about it?