No, it won't.
Garbage in, garbage out.
The LLM-based AI systems are very good at matching patterns in text and extending those patterns in very complex ways. Current designs have absolutely no concept of whether those texts have anything in them that correctly represents the real world.
They cannot tell what is true and what is false.
LLM programs can only be as good as the curated texts using for training them, and the quality of the responses deteriorates quickly if the responses are fed back into the program for revision queries. The programs are prone to generating false responses (e.g. "Hallucinations").
Just because a response text is logical, does not mean it is true or correct.
Garbage Algorithm, Garbage Out.
"AI" with trash for training data combines the worst of both worlds.
It will seem to be a God eventually