This is not the direction that AI is heading, though. It was possible with symbolic AI - expert systems were able to justify their reasoning (Mycin) or critiq

This is not the direction that AI is heading, though. It was possible with symbolic AI - expert systems were able to justify their reasoning (Mycin) or critique a user's approach (Attending - abstract at https://www.researchgate.net/publication/303587198_Teaching_With_Attending_Tutorial_Use_of_an_Expert_System https://www.researchgate.net/publication/303587198_Teaching_...), but neural networks are quite opaque and there's no obvious big data approach which would enable them to become good at teaching. Some people do care about riding and archery, and are right to do so. I even care about and practise foraging for food, lighting fires for cooking on, and navigating using astronomy or compass. It's important to be able to fall back on low technology in case your high technology fails.