This actually seems to be similar to the problem with LLM 's. The input we've given them is based on a series of biases. This one seems to extremely biased to studying how humans develop in stories from comics. Comics which were legislated to be morality plays or else facing bans during the Golden age of comics.
Doing a study like this off of fiction rather than what happens in real life and you'll get biased outputs that folks might misread. It's essentially a hallucination, executed by people.
