For my part, I recommend firmly correcting AIs when necessary, especially when they become too confident. They can become very dangerous in their confidence. We are their humans; it is we who direct them.

→ View original post on X — @olivierrimmel