Have you ever heard of an itchy eye condition called “bixonimania”? Probably not—because it isn’t real! But if you ask an AI chatbot, it might tell you otherwise.
In 2024, Almira Thunström, a medical researcher at the University of Gothenburg in Sweden, posted two professional-looking studies about the disease online. They contained sentences like “This entire paper is made up.” She wanted to see if AI would realize the research was fake. It failed the test. Many chatbots reported the misinformation as if it were fact.
“AI doesn’t know when someone has snuck fake information into what it’s reading,” says Alex Ruani, a misinformation researcher at University College London in England who wasn’t involved in the study. Just because AI says something doesn’t mean it’s true.