Absurd. It is not the software doing the evaluating, it is the bias of the code writer.
“It is not the software doing the evaluating, it is the bias of the code writer.”
That’s not how self-learning AI networks work. What you are claiming pertains to classic deterministic code.