Raw LLM Responses

These experiments with LLMs (what some people call AI) harming people when there's a conflict of interest (or risk of them being shut down) sound like they were influenced to get the results they wanted to get fame/virality. I'm gonna look more into it but they sound faked to some degree. I'd also like to mention that fear mongering and grandiose statements about LLMs made by LLM researchers/investors could be influenced by the financial interests of these people. Propping up LLMs as being very dangerous could lead to more funding, and more status for the researchers, while they're just tools that predict words through statistical analysis. This is not to say that LLMs don't pose any threats to our society. Global warming and wasting of resources are big threats, though they are also often sensationalized ("a single AI generated picture wastes enough water to feed a village for a day!" when in truth it's a small portion of a liter, still awful but not nearly as much as some claim). A bigger threat is the data that feeds LLMs itself being "AI generated", creating a loop that not only reinforces already existing biases but also leading to a state where they only regurgitate the same statements. Another threat is people not realizing that these things can not create new results and becoming too dependent on them.