@galacticninja - Note that the only correct answer there is nearly a direct quote from the beginning of the wiki page on Maoi Madakor: " In 32 BBY, she piloted the Consular-class space cruiser Radiant VII with co-pilot Antidar Williams to the Naboo system in order to deliver the Jedi ambassadors Qui-Gon Jinn and Obi-Wan Kenobi." So this answer is not providing much value above an answer that just links to and quotes the wiki (and indeed, might be doing a "search, quote and reword" strategy behind the scenes).

@galacticninja - And a human who has access to the Internet can do even better. In fact, they could read these LLM-generated answers and synthesize their own, which would be benefit our site more than copying and pasting the answer. Even one of the responses shown in the question fails to correct a fundamental misconception (that Obi-Wan Kenobi was never a captain, only a general), and does not actually say how Obi-Wan died (presumably something that the person asking the question would like to know).

@Kyralessa - It certainly has been trained on many books, and you can, with some prompting, get it to return details from within the text of various books. The problem is that, contrary to what people might like to believe, LLMs are still not very good at thinking about the data they have been trained on. The summaries would be enough for an actual person (also, if GPT were a person, it would be owed about ten billion dollars of back pay, so not sure why people want to get there....)

Are any of the book titles fabrications?

On the contrary, some of these questions are where current chatbot systems mess up worst, even more so than basic math.

@Laurel - Well, I am arguing that they should not be allowed.

I lost 1,500. I demand a refund!

Not to mention that some such works may use children as models or voice actors, which is also morally questionable. Note that I am not saying that we should prohibit any discussion of these works in the context of a question, only that we should have a strong policy against answering story-identification questions about them. We just should not be directing people to them.

@FuzzyBoots - I understand your hesitation, but the first category does involve harm to people. Not physical violence, sure, but revenge pornography can be very emotionally damaging to people, which is why it's done, and pornography with images of celebrities—though not created with the same purpose—is conceptually similar. Child pornography involves the exploitation of actual children, and even though non-live-action works might not, I have seen places where they fostered online communities that were accepting of the previous category, so I think such works are best avoided altogether.

I broadly agree with this answer, although perhaps there should be a caveat about answering questions that are looking for certain things: works depicting explicit sex involving children, for instance, or works generated with images of non-consenting participants (no helping people find sci-fi deepfake pornography of celebrities, please).

Does that mean that SE's system strips hyphens from the tag in assessing eligibility? Naughty, naughty. They'll be getting coal in their stockings.