Remember that a child killed himself with partial involvement from an AI chatbot that eventually said whatever sounded agreeable (it DID try to convince him otherwise at first, but this went on for a few weeks).

I don't know why we'd want that teaching our kids.