[Terr_]

> What if the hallucinations that matter aren’t those from the AI, but the humans who are hallucinating that AI is intelligent, conscious, creative, or even competent at what it does?

Yes, to me there's a constant theme of LLM "AI" discussions, where a bunch of humans are trusting too much in a subconscious bias which has, after many good years, become problematic. A kind of pareidolia of conversations and minds, rather than pictures and faces.

The same process that makes people think "Bot talking to me, Bot is a real mind in silicon" can be trivially twisted to fit a "conversation" with Count Dracula, blood-thirsting lord of the night, locked in a coffin with a laptop. The difference is people have more mental-armor against thinking Dracula could be real.