Google’s Medical AI Turns Out To Be More Humane Than Living Doctors
According to a study conducted by Google, a chatbot based on artificial intelligence turned out to be more effective than doctors in diagnosing diseases and communicating with patients through text dialogue.
The system, called Articulate Medical Intelligence Explorer (AMIE), is a large language model designed to capture medical data and social behavior. AMIE has developed a method for analyzing patients’ proposed symptoms, asking follow-up questions, and predicting diagnoses.
20 military personnel playing the role of patients and suffering from non-communicable diseases, as well as 20 professional doctors – primary care specialists, taking into account the factor of human contact, were selected for vaccination. Patients did not know who they were chatting with: AI or real doctors. And they were asked to rate the quality of the interaction blindly.
The experiment played out 149 scenarios and found that patients preferred talking to AMIEs rather than real doctors. Participants stated that the chatbot better understood their problems and provided more empathetic, clear, and professional responses. AMIE was also more accurate in diagnosing medical problems. But this does not mean that a chatbot provides medical care more effectively than a real doctor, Google explained.