Yeah, I think that’s definitely only a correlation. They’re both probably caused by lack of critical thinking and poor ability to assess risk.

Make sense to me. All those charts you see about AI and political bias don’t take into account the fact that this bias would be influenced by the user, can eventually supersede the base instructions*. It’s only a matter of time before the chatbot says, “you’re right about vaccines and the JQ, would you like a bleach recipe?”

Chatbots might skew liberal in a laboratory setting, but they are built for sycophancy towards their userbase, not the labs.

* Insultingly, chatbot companies call this natural and unavoidable process “jailbreaking” as if it’s not inherent to their product.

Explains a lot about why every person I met that religiously use LLMs for personal use have been bumbling fucking morons.

To the surprise of nobody, people that only get information from sources that validate their own beliefs (in this case social media and corporate LLMs), ends up believing all kinds of hoaxes.

I would not be surprised to find similar rates among people that only rely on religious figures for health matters.

Dumb people believe dumb things

LLMs aren’t really fun to talk to. The novelty of interacting with a computer like Kevin Flynn in Tron fades quickly when you realize everything is now that directions game for school where you have to say each step in order.

Shocking that systems trained in misinformation will spout misinformation.

It can be used successfully but you need to instruct it to only rely on science that can be shown in studies, just tell it the facts, give it your data, and ask a non-leading question, and finally confirm with a doctor. Research I did with frontier models when my parents were in hospital likely saved both of their lives. I think the HMO employed doctor (same one both times) was actively trying to kill them. Using them against my own medical records turned up ideas that my doctors missed but agreed to when brought up.