Study Assesses the Accuracy of Medical Advice from AI Chatbots

Sat 10th May, 2025

In a recent study, a team of researchers from various institutions in the U.K. and the U.S. assessed the reliability of medical advice provided by AI chatbots, including popular models like ChatGPT. The findings, published on the arXiv preprint server, highlight the challenges of obtaining accurate health information from these artificial intelligence systems.

The researchers engaged 1,298 volunteers, asking them to seek medical advice from AI chatbots or rely on their usual sources, such as internet searches or personal knowledge. The goal was to evaluate how well the chatbot responses compared with information from other online medical resources.

As the demand for immediate medical guidance grows, many individuals are turning to chatbots as a convenient alternative to traditional consultations. While previous studies have demonstrated that AI applications can excel in clinical examinations, this research aimed to explore how effectively these capabilities translate into real-world interactions with users.

The volunteers were randomly assigned to either the chatbot group or a control group that used their preferred methods for seeking medical advice. The interaction between the volunteers and the chatbots was recorded and later analyzed by the research team.

One of the significant findings was that many volunteers provided incomplete information during their queries, hindering the chatbots' ability to accurately assess the medical situations presented. This breakdown in communication often led to suboptimal advice from the AI systems.

When the researchers evaluated the suggestions made by the chatbots against those provided by other sources, including online medical sites and the participants' own instincts, they discovered that the chatbot advice was occasionally comparable but often inferior. In numerous cases, the use of chatbots resulted in volunteers misidentifying their health issues or downplaying the seriousness of their conditions.

The study concluded that while AI chatbots can offer some level of assistance, reliance on them for medical advice may not always be prudent. The authors strongly recommend that individuals seeking health information opt for more traditional and trusted sources to ensure a more accurate understanding of their medical needs.

This research underscores the complexities of integrating AI into healthcare and the importance of clear communication in medical inquiries. As technology continues to evolve, further studies will be essential to determine the best ways to utilize AI in supporting health outcomes without compromising patient safety.


More Quick Read Articles »