Post #2844660

IA slop. Tested, not approved… To test how easy it is to poison the information spewed out by AI-powered search results and chatbots, medical researcher Almira Osmanovic Thunström and her team uploaded two fake studies to a preprint server in early 2024. They included plenty of red flags: the results even say “this entire paper is made up”. But soon after, systems such as Copilot and ChatGPT were telling users all about ‘bixonimania’ as if it were real…