Ok. Let's pivot to: if the captcha is properly solved without one fault, then it's a bot :) Humans verify themselves by mouse movements, by errors with hard to

Ok. Let's pivot to: if the captcha is properly solved without one fault, then it's a bot :) Humans verify themselves by mouse movements, by errors with hard to read chars, or, and I think that's will do it, generating something unrealistic with a AI (like a sweet red dog with chicken feet in a cooking pot) and let the human guess what's wrong in the pictcha.. Or some sort of.. questions with a double-meaning, and only if you utilize the understanding of the world outside, that ones can be solved. Like an old apple in a sitting bank and then present the pictures of an old apple computer within a money bank. A real, dry apple between a sitting bank's lattice.. Some crazy stuff can be done with AI, why not to utilize it?