Thank you for your detailed and thoughtful response, Boris. Let me point out that this was by no means intended to be a comprehensive experiment or broad assessment of AI in general, but rather a description of what we experienced from a single case of the sort of query I would expect people might be interested in posing to AI chatbots. Obviously, this particular query is very narrow and of interest to a limited number of people, but I think it's typical of the kind of questions some people might ask.

Your points are all valid, of course. Certainly, there are many applications for which AI is very well suited. And here's an example of one on which it does IMO a poor job, yielding some results that are clearly not accurate. I'd rather the AI simply said "I don't know" or returned a more limited response, instead of giving the user information that is simply wrong. How is that not a "use at your own risk" situation?