MouthBQ98 said:
Institutionalized stupidity. We've flooded media with so much intersectionalist identity politics garbage over the last few years, of course the AI are going to pick it up and run with it. They see data to incorporate and don't understand the context or politics necessarily.
This is directly related to the prompts and behind the scenes code that the model uses to process information and not so much whatever is rampant on the internet or in media.
Focused LLMs like this generally aren't exposed to the general internet to "learn" and pick up a bias. They're sandboxed for a specific task. This is 100% programmed in.