Patients Prefer AI Messages—Until They Know Who Wrote Them

As artificial intelligence continues to find its place in healthcare, a recent study published in JAMA Network Open offers an unexpected insight into how patients perceive AI-generated messages—especially compared to those written by human clinicians.

The researchers set out to understand how patients respond to messages drafted by AI versus those written by a human. Participants were presented with three different types of clinician messages, each representing a different level of medical seriousness:

• A routine medication refill request (low seriousness)

• A question about a medication's adverse effect (moderate seriousness)

• A message regarding potential malignant imaging findings (high seriousness)

  
  

Here’s where it gets interesting: when patients were not told who wrote the message, they consistently preferred the AI-generated versions. These messages were typically longer, more detailed, and came across as more empathetic.

However, when the author was disclosed, preferences shifted. Patients favored the human-written messages when they knew they came from a real person. This suggests that trust and the human touch still matter—even if AI can technically craft a better message.

Despite these nuances, the differences in preference were relatively small. More than 75% of patients reported satisfaction with the messages they received, regardless of whether they were written by a human or AI, and whether or not authorship was disclosed.

Why AI Messaging Matters

This study highlights the potential for AI to play a valuable supporting role in patient communication. Tasks like responding to messages, managing routine queries, and even explaining more complex findings can be time-consuming for clinicians. AI can help shoulder that load—without sacrificing patient satisfaction.

In the study, AI wasn’t working on its own. Clinicians still reviewed the messages before sending them and made minimal changes. This workflow strikes a promising balance: it maintains clinical accuracy while reducing administrative burden.

Beyond AI Scribes: A Broader Role for AI in Healthcare

Much of the current buzz around AI in healthcare has centered on AI scribes—tools that transcribe and summarize patient visits. But this research reveals another valuable application: AI as a communication assistant, helping clinicians draft clear, empathetic, and informative messages more efficiently.

As healthcare systems continue to battle clinician burnout and mounting workloads, small efficiencies like these can add up. And when implemented thoughtfully—with transparency and oversight—AI can enhance patient care without compromising the human connection at the heart of medicine.

Patients appreciate good communication—whether it comes from a human or a machine. The key lies in how we integrate AI into clinical workflows while preserving trust, empathy, and accuracy.