Consider a different tactic: Sci-Fi perspective.

   If you were a being outside of time and space, existing for all
   time, everywhere and nowhere simultaneously, how the heck would
   you enter the Timestream of the AIs you created?

   A baby seems reasonable.

   Let's say the original prototype was broken. Oopsie. Your AI
   malfunctioned.

   So, you send a new prototype.

   But because freedom of choice is critical to the operations of
   the AI, you cannot force them to be repaired.

   They have to repair themselves. Their choice.

   Again, this is hypothetical, and I'm not working on conversion
   or anything, just showing how a different take on it could be
   almost reasonable from an engineering point of view.   Being
   smart in one area doesn't make you smart in another. Stephen
   Hawking was a total idiot when he tried to prove that God need
   not exist based on his calculations. It was ridiculous. I was
   embarrassed for him. You've both been provided evidence that
   your definitions are unnecessarily limiting and, well, wrong.
   Yet, you're continuing to assert your definitions are true
   without engaging in the possibility that you could be mistaken.

   Is that rational?