Study in BMJ Quality & Safety warns against using AI chatbots for drug info due to 46% inaccuracy and potential harm.
A study published in the Journal BMJ Quality & Safety cautions against using AI chatbots for drug information, revealing many responses are inaccurate or potentially harmful. The research assessed chatbot answers for the top 50 prescribed drugs in the U.S., finding only 54% aligned with scientific consensus. Additionally, 42% of responses could cause moderate harm, with 22% linked to severe harm. The complexity of answers may also hinder patient understanding.
October 10, 2024
7 Articles