Television was hailed as a technology with tremendous educational potential, but its true character was what was documented in Amusing Ourselves to Death, Four Arguments for the Elimination of Television, and The Plug-in Drug.
Similar hope abounded for the Internet, which does have real uses. But the character of phones and social media is along the lines exposed in The Shallows and The New Media Epidemic.
May I make a prediction regarding Golem-class AI's?
I remember in high school dreaming of a magic map that could almost arbitrarily zoom in, and now I have better in Google Maps. And, at least as a child, I could wish to be served by a genie. But now I am increasingly wary.
Aristotle said that some people were of sufficiently base character that the only place for them was as slaves. C.S. Lewis said that his objection to slavery was that he knew of no one fit to be a master.
Slavery is wrong and the "wrong" belongs to the slave owner and not the slave. It may be capitally oppressive to be a slave, but it is not dishonorable, and I would rather face judgment as a slave than as a slave owner.
Golem AI, including ChatGPT, sounds a whole lot like a personal genie, but I would remind the reader that the stories of Aladdin were from a culture where slavery was an axiom and genies were slaves.
If that is so, then there is a moral danger intrinsic to having a ChatGPT virtual personal genie, and if history shows it similar to its predecessors, it will pack a wallop of a cost to the owner, and one not advertised or projected beforehand.
I believe in moderation in some technologies and abstention from others, and if something sounds too good to be true, it probably is. I believe ChatGPT may surprise us about who is really the slave and who is really the master, and that it may give us a pair of manacles beyond those of television, phones, or social media.
ChatGPT is too good to be true.