A new study finds that individuals increasingly trust medical advice from artificial intelligence, even when the information is often inaccurate. Researchers from the Massachusetts Institute of Technology published their findings in the New England Journal of Medicine, revealing that participants preferred AI-generated medical responses over those from human doctors.
In the study, 300 participants, including both medical experts and laypeople, were asked to evaluate medical advice provided by a medical doctor, an online healthcare platform, or AI models like ChatGPT. Surprisingly, many rated the AI-generated responses as more accurate, valid, trustworthy, and complete. The study indicated that neither group could reliably distinguish between AI and human responses.
Participants were also exposed to AI-generated advice that had low accuracy, which they were unaware of. The findings revealed a concerning trend: individuals not only perceived these low-accuracy responses as valid but also expressed a strong inclination to follow this potentially harmful advice. This could lead to unnecessary medical consultations and treatments.
Dr. Darren Lebl, research service chief of spine surgery for the Hospital for Special Surgery in New York, highlighted the dangers of relying on AI for medical advice. He noted that many AI-generated recommendations lack a scientific basis, stating, “About a quarter of them were made up.”
Notably, there have been documented cases where AI has given dangerous advice. An incident involved a 35-year-old man from Morocco, who sought emergency medical assistance after a chatbot advised him to wrap rubber bands around his hemorrhoids. Another alarming case involved a 60-year-old man who was hospitalized for three weeks after consuming sodium bromide, a substance used for pool sanitation, based on suggestions from ChatGPT.
Further emphasizing public sentiment, a survey conducted by Censuswide found that approximately 40 percent of respondents reported trusting medical advice from AI bots. This growing reliance on artificial intelligence for health-related guidance raises significant ethical and safety concerns.
As AI technology continues to evolve, the implications for medical practice and public health are profound. The challenge lies in ensuring that individuals receive accurate, evidence-based information while also navigating the allure of AI’s accessibility and convenience.
