November 16, 2026

Why Clinicians Hesitate to Trust AI, and What Could Change That

happy-young-african-american-assistant-uniform-looking-mature-woman.jpg

Clinicians may hesitate to trust AI recommendations when they cannot see the evidence behind them or understand how confident the system is in its conclusions. Building clinician trust in AI requires explainability, confidence scoring and a clear role for human judgment, particularly when decisions affect patient care.

What would make clinicians more willing to act on an AI recommendation? And how can healthcare organizations use AI to improve an entire process rather than simply automate individual tasks?

These are among the questions Zyter CEO Sundar Subramanian explored with John Marchica, CEO of Darwin Research Group and host of Healthcare Rounds. Their conversation explored the practical challenges of building trust in healthcare AI, redesigning workflows and measuring whether these efforts are making a difference.

Listen to the full Healthcare Rounds episode

Why clinicians need to understand AI recommendations

During the conversation, Sundar examines why even an accurate AI recommendation may be questioned or rejected.

A clinician needs to understand the reasoning behind a recommendation, how much confidence to place in it and when additional review is necessary.

Research into AI-assisted diagnoses has shown how transparency and confidence scoring can reduce the likelihood of recommendations being overridden.

For Sundar, this raises an important question about how AI should operate in clinical settings: how can systems recognize uncertainty and bring a clinician into the process when additional judgment is needed?

Rethinking the work AI is supposed to improve

Sundar also challenges healthcare leaders to look beyond individual AI applications and consider how an entire process could work differently.

Prior authorization offers a practical example. Information must be gathered, clinical documentation reviewed and decisions communicated across multiple parties. AI can assist with these steps, but improving one task doesn't necessarily resolve delays elsewhere in the process.

Sundar discusses how AI orchestration can help coordinate agents, systems, business rules and people across the workflow, with human involvement where appropriate.

Looking ahead, the conversation highlights three areas that could indicate whether these efforts are delivering results:

  • Faster processes: Reducing delays across the full workflow, including handoffs between systems and teams.
  • Better patient outcomes: Connecting operational improvements to measurable changes in care.
  • Greater clinician confidence: Giving healthcare professionals the information and oversight needed to use AI recommendations appropriately.

These considerations bring the discussion back to a question healthcare leaders will increasingly need to answer: Is AI improving how care and operations actually work?

Listen to the full conversation

Hear Sundar and John discuss clinician trust in AI, the role of human judgment, prior authorization and how AI orchestration could reshape healthcare operations.

Listen to the full episode

Related Articles