The Fake Disease That AI Confirmed as Real

The bixonimania experiment reframes where the AI misinformation threat actually lives. The research team did not exploit a model's internal weaknesses or craft adversarial prompts — they exploited the gap between what gets published online and what gets formally vetted. Two fake studies, complete with invented authors and a fictitious funding body named after a fantasy novel, entered the information ecosystem through the same channels legitimate research does. The , but the chatbot outputs they generated did not disappear on the same schedule. Publication was…