The study compares diagnoses made by AI and clinicians and shows how AI could be integrated into health care.

ORONO, Maine — A new study from the University of Maine shows how artificial intelligence could be a helpful tool for doctors diagnosing medical conditions.

Dr. C. Matt Graham, a professor at UMaine, evaluated more than 7,000 medical queries to compare how clinicians and AI diagnose serious health conditions. He found that the AI responses were more accurate and professional than the human responses.

The AI bots were highly factual with a consistent structure and tone. But the AI often struggles with follow-up and “how” and “why” related questions.

The main difference between the AI and doctor responses was emotion. The AI lacked empathy and used more clinical terms, while the doctors offered more compassion, which Graham said is critical to patient care.

“AI really does need to be complimentary to the human clinicians,” he said. “The goal of this research was not in any way, shape, or form to demonstrate that AI can be used to replace human clinicians in this context. I see AI as simply making the human clinician better at their job.”

There are a few ways that Graham said AI could help lighten the workload of health care workers in Maine, especially as many hospitals struggle with staffing and rising costs. He said using AI could help streamline documentation and data analysis. As the study shows, AI could also help support doctors make initial assessments and diagnoses.

“It shows promise in reducing clinician burnout by handling repetitive patient communications, and then it can also be used for patient education tools,” Graham said. “AI can be used for training and simulation, for medical training and diagnostics with patient interaction scenarios.”

Graham added there are also some risks and challenges when it comes to integrating AI. He said most AI tools are trained on limited populations, so there needs to be more inclusive data, human oversight, and regulations to avoid inequities in health care and keep patient safety at the forefront.