The problem with Artificial "Intelligence" is that it really has no intelligence at all. Intelligence requires understanding, and AI doesn't understand either the data fed into it or the responses it gives.
Yet because these tools output confident, plausible-sounding answers with a professional tone (which may even be correct a majority of the time), they give a strong illusion of being reliable.
What will be the result of the current push of GPT AI into the mainstream? If people start relying on it for things like summarizing articles and scientific papers, how many wrong conclusions will be reached as a result? God help us if doctors and engineers start making critical decisions based on generative AI answers.