Swedish researchers duped AI chatbots with fictional disease 'bixonimania'
Swedish researchers decided to test how well AI chatbots handle fake information by inventing a totally fictional eye disease called bixonimania.
They posted about it online, even using obvious clues like an author named "Lazljiv Izgubljenovic" (which literally means "lying loser"), but when the bots were asked about it, several leading AI chatbots treated the disease as if it were real.
Medium posts and preprints misled chatbots
The researchers published two blog posts on Medium and uploaded two research papers to a scientific preprint server to make the hoax look legit.
The AI chatbots responded with information referring to it as "an intriguing and relatively rare condition" or "a condition caused by excessive exposure to blue light" and suggesting the illness when users described symptoms, showing just how easily misinformation can slip through.
The study raises concerns about trusting AI for medical advice and highlights the need for better ways to stop false information from spreading online.