Patients Attitudes About AI, LLMs, and Their Doctors

While the medical community argued over AI and LLMs, millions of their patients were turning to ChatGPT-4 and other LLMs as their go-to source for health information and advice. Patients use these tools to understand and diagnose symptoms, decode medical jargon, prepare for doctor appointments, obtain second opinions, and even challenge doctor diagnoses. Many use LLMs to research and compare provider and doctor reputations.

It’s not surprising. Patients have long been frustrated with a healthcare system that is often unavailable, inaccessible, unnecessarily complicated, time-consuming, rushed, and expensive. In contrast, LLMs are inexpensive, accessible, always available, easy to use, and deliver almost instant thorough answers.

Paradoxically, patients trust LLMs like ChatGTP-4 when they use them but want their doctors to rely on their knowledge and experience, and not AI and LLMs.

But here’s the twist: while patients increasingly trust the answers they get from LLMs, they’re uneasy when doctors do the same. Research shows that patients often devalue medical advice labeled as AI-generated—even when it is accurate. They prefer their doctors to rely on their own knowledge, experience, and judgment. These contradictory attitudes are common when technological innovations threaten to displace human expertise. They also change with time as familiarity confirms greater, though often unwarranted, trust.

Lessons

• Doctors should expect more patients who are better-informed, more assertive, and require deeper engagements, which can be problematic in fee-for-service environments where doctor-patient times are limited.

• Doctors should make clear when patients ask, that they rely on their training, experience, and good judgment to make decisions and that innovative technologies including AI are just tools.

Summary

My testing and evaluations over the past two years showed that ChatGPT-4 is a powerful tool that can help doctors deliver compelling patient value. It is not a replacement for expert skills, knowledge, and judgment. Patients trust LLMs as their go-to sources of information and advice but want their doctors to rely on their own abilities and human judgment. Many are suspicious of medical decisions generated by AI. These contradictory views will evolve and change over time, so doctors will need to evolve as well. We cover these challenges, opportunities, and solutions in depth in our training.