The artificial intelligence chatbot ChatGPT has garnered attention for its proficiency in medical diagnostics.
A study conducted in the United States found that ChatGPT achieved a 72 percent success rate in diagnosing illnesses and selecting appropriate treatments.
Experts from Mass General Brigham, a US-based healthcare research organization, suggest that ChatGPT could become an invaluable tool for physicians in the future.
Researchers uploaded detailed health data from 36 patients into ChatGPT and requested the AI to suggest diagnoses and treatments.
Upon comparing ChatGPT’s recommendations with actual case outcomes, it was revealed that the AI had an average success rate of 72 percent.
Notably, it attained a 77 percent accuracy rate in determining final diagnoses. However, ChatGPT faced challenges in differentiating between similar diseases, with its success in making differential diagnoses dropping to 60 percent. The AI’s accuracy in choosing the correct treatment was 68 percent.
Scientists noted that ChatGPT is also capable of making accurate decisions in emergency situations. They compared the AI’s performance to that of a trainee doctor fresh out of medical school, highlighting that while ChatGPT’s success is commendable, it does not possess the expertise and decision-making capabilities of human physicians.
The findings of this research were published in the Internet Journal of Medical Research. Researchers highlighted ChatGPT’s advantages in summarizing patient histories and focusing on minor details, suggesting that AI chatbots like ChatGPT could assist doctors in diagnosing patients.
You may also like this content
-
-
-
-
-
-
-
-
-
-
