AI models hallucinate because they are probabilistic engines designed to predict the next likely word. They don't know what "truth" is, they just say what is "plausible-looking" based on training data. Some numerical inaccuracies in the model may also contribute to "hallucinations".

Humans hallucinate under sensory deprivation because, I think, our brains never evolved to respond reasonably to full sensory deprivation (it basically almost never happens in real life to healthy individuals who leave offspring), i.e. this kind of scenario was never "tested" by evolution and we get results that make no sense (random garbage).

Of course, yes. Yes. It is also sad when it is hungry so best feed it with its preferred food, which is dollars. Loves to eat dollars.

So, another religion ?

That's an interesting conjecture. Many of the attributes of AI that people object to (hallucination, sycophancy, psychosis, plagiarism, etc.) originate from human behavior. We see ourselves in the system and, in a way, are the ghost in the machine. So yeah, it could be that AI systems need socialization, and that may be a job for humans in the future. I'm now waiting for the first Dr. Susan Calvin.