Why is the ability to ask questions so rare among other animals, relative to humans? There has only been one recorded case of a non-human animal that could ask questions(a captive african grey parrot), whereas at least 20% of humans are capable of forming questions.

Are we going to suddenly hit a huge hill in machine learning where everybody struggles to figure out how to make a machine that is capable of asking questions? It's already a known issue(this falls under something called "scalable oversight"), but it's not known how difficult the issue is.