What responsible AI should look like in 2026

Overview

In a recent Medical Economics article, Dr. Brian Covino, chief medical officer for Cohere Health, makes the case for a new standard in AI-driven utilization management–one where automation accelerates decisions without compromising the clinical judgment, transparency, and physician oversight that patient care requires.

As federal regulations and industry commitments raise the bar on speed, visibility, and payer-provider collaboration in prior authorization, Dr. Covino argues that modernization and clinical intelligence aren’t competing priorities. Responsible AI in prior authorization must be built on domain-specific, clinically trained models, policy-aligned recommendations, and explainable rationale–earning physician trust through rigorous validation, governance, and clear accountability, not by replacing clinical expertise altogether. 

Responsible AI becomes a clinical imperative–supporting judgement, strengthening trust, and enabling appropriate care to move forward safely and efficiently.

Key themes covered:

1. Why physician skepticism of AI in utilization management persists
2. What differentiates responsible, clinically trained AI from generic automation
3. How explainability and oversight build lasting physician trust
4. Where clinically trained AI is streamlining prior authorization and downstream payment operations today

Read the original article