Whithout a clear definition of intelligence, such a discussion is somewhat pointless. The closest thing I would use to describe artificial “intelligence” is: Mimicking human thought and decision processes, without the necessity of being identical.
Secondly, regarding your first paragraph:
Humans excel at providing wrong information. Sometimes they are right. In that regard there are similarities between an LLM and a human.
By the way: LLMs are part of the field of AI. But AI consists of a plethora of methods and algorithms, where LLMs are just a tiny fraction that is currently very popular.
haloduder@thelemmy.club 6 days ago
I agree. AI would be way more useful if it gave some kind of indicator of accuracy, or even just saying “I don’t know.”
Right now, AI only doesn’t have an answer when the goons who programmed it told it not to give one.