We leave it to the AI.

We have gone to a lot of trouble to try to engineer the prompt to make it clear to Gemini that it should take a "non-partisan and unbiased" view in all the analysis. This is an attempt to get away from any person's opinion, including ours.

Obviously, whether you think it achieving that ideal is in the eye of the beholder :-) But it is certainly less biased than most mainstream media, and social network echo chambers.

So the bias is left to the AI's training set?