But there still remains the possibility of error/hallucination. Do you bet on a future where AI will never make any mistakes whatsoever?

I am often told humans make mistakes, too. But this is an elementary logical fallacy because the type, frequency and severity of the problems is never the same.

Also, humans are more likely to do as told, whereas AI often takes liberties (i compiled a book from a github repo and it did things on its own in the book though i had explicitly forbidden that).