How well can AI diagnose patients compared to human clinicians?

A University of Maine study compared how well artificial intelligence models and human clinicians handled complex or sensitive medical cases. The findings outlined where the technology showed promise and what limitations need to be addressed before AI is unleashed on patients.

With growing provider shortages, especially in rural states like Maine, the study highlights how AI could help clinicians. Lessons from past integration of technology, which at times failed to enhance care delivery, offer valuable guidance for AI developers as they work to integrate the technology into clinical practice.

The study authored by C. Matt Graham, associate professor of information systems and security management at the Maine Business School, analyzed 7,000+ medical queries from the U.S. and Australia, comparing AI-generated responses with those of human providers. The verdict? AI shows potential as a clinical support tool, especially in data-heavy or procedural scenarios, but struggles with emotional nuance, cultural context and consistency over time.

“Technology is only one part of the solution,” Graham says. “We need regulatory standards, human oversight and inclusive datasets. Right now, most AI tools are trained on limited populations. If we’re not careful, we risk building systems that reflect and even magnify existing inequalities.”

Discussions about how to make the most of AI’s possibilities will take center stage at the Maine AI Conference this Friday. We’re bringing together tech and business leaders, educators, policymakers and researchers to explore how Maine can lead the way in responsible AI.

Read the full story on UMaine News.

Contact: Erin Miller, erin.miller@maine.edu