I'm not sure a higher level "intelligence" (which some folks think AI is moving towards) should be overly-reliant on human "intelligence", lest it inherit flaws

I'm not sure a higher level "intelligence" (which some folks think AI is moving towards) should be overly-reliant on human "intelligence", lest it inherit flaws which may outnumber benefits. (Humans believe a variety of outlandish things, such as "Q-Anon has the real facts", etc.)