In regards to ChatGPT: My concern is the output not getting verified enough by knowledgeable experts. I've played with it a little. It is impressive, but at the same time it becomes harder to distinguish 'accuracy' from the 'almost correct', ánd to detect the error in between.