You're completely wrong.

"better" is an objective evaluation that you can do in a test, not in an interview with an AI.

And AIs always have human bias encoded, because it's trained on human data. That's a well known problem with no absolute solution.