From a far off perspective, can we really assume any model to be free of bias (political or otherwise)? I don't know - but would like to see - how the political opinions of reddit users and blog authors are reflected in that of AI, as AI is trained on human (biased) sources.