With millions of Americans increasingly turning to AI chatbots for health advice, two new studies suggest that reliance may be riskier than many realize.
Recent research found that leading AI tools—including ChatGPT, Gemini, and others—struggled to provide accurate medical information. In one study, chatbots answered just over 50% of health questions correctly, with roughly 20% of incorrect responses deemed potentially dangerous if followed.
At the same time, new data from the West Health-Gallup Center on Healthcare in America shows that 1 in 4 Americans now use AI for health information—and about 14 million have skipped seeing a healthcare provider based on chatbot advice.
Ioannis Koutroulis, the dean of MD Programs at the George Washington University School of Medicine & Health Sciences and associate professor of pediatrics and of emergency medicine is available to discuss:
- Why AI chatbots struggle with medical reasoning and diagnosis
- The growing number of Americans substituting AI for professional care
- Risks of “confidently wrong” health advice
- How healthcare systems are adapting to increased AI use
To schedule an interview, please contact Katelyn Deckelbaum, katelyn [dot] deckelbaum
gwu [dot] edu (katelyn[dot]deckelbaum[at]gwu[dot]edu).