AI chatbots provide poor answers to medical questions half the time, study finds

“The audited chatbots performed poorly when answering questions in misinformation-prone health and medical fields,” the researchers concluded. “Continued deployment without public education and oversight risks amplifying misinformation.”