Research shows that popular AI chatbots provide inaccurate or unsafe health advice nearly 50% of the time [2].
This trend is concerning because a growing number of people use these tools to bypass physician shortages and long wait times for specialist care. Without proper guidance, users may rely on confident but incorrect medical information that lacks expert verification.
One in three American adults said they have already used an AI tool for health-related questions [3]. Other data indicates about one in six U.S. adults specifically turn to AI chatbots for self-diagnosis [5].
An audit of popular free AI chatbots found that these systems often deliver misleading information and weak citations [4]. The lead author of the audit said the tools often sound confident while delivering advice that may be unsafe without expert guidance [4].
To mitigate these risks, Dr. Sudheesha Perera and Dr. Murali Doraiswamy have provided guidance on how to obtain more reliable answers. Perera said AI can help fill a real gap left by physician shortages and long waits for specialist care if it is used wisely [1].
Users are encouraged to use specific prompting strategies to improve the quality of the output. This approach aims to ensure the AI provides more accurate information and better citations, reducing the risk of following unsafe medical suggestions [1], [4].
“AI chatbots provide inaccurate or unsafe health advice nearly 50% of the time.”
The gap between the accessibility of AI and its medical accuracy creates a significant public health risk. As U.S. healthcare costs rise and provider shortages persist, patients are increasingly incentivized to use unverified tools for self-diagnosis. The high rate of misinformation suggests that AI is currently a supplement to, not a replacement for, professional medical consultation.
