University of Utah Researchers Unveil Four-Tier Framework for AI Automation in Psychotherapy

A new study from the University of Utah categorizes AI in therapy into four levels, from scripted tools to autonomous agents, focusing on risks and clinical utility.

By: AXL Media

Published: Apr 7, 2026, 6:05 AM EDT

Source: Information for this report was sourced from the University of Utah

A New Paradigm for Clinical Collaboration

Psychotherapy has historically been defined as a deeply human process of healing through spoken words. However, the rapid ascent of Large Language Models (LLMs) is fundamentally shifting this landscape. Researchers from the University of Utah’s College of Engineering, School of Medicine, and College of Education have published a landmark study titled "A Framework for Automation in Psychotherapy" to provide a roadmap for this transition. Lead author Zac Imel, a professor of educational psychology, emphasizes that the goal is not to replace human experts but to identify how technology can support them in delivering more effective, data-driven care.

The Four Categories of Automation

To simplify the complex role of AI in the clinic, the team established four distinct categories along a continuum of risk and autonomy:

Category A: Scripted Systems – Chatbots that deliver prewritten, human-vetted coping tips following fixed decision trees.

Category B: AI Evaluates Therapists – Systems that review recorded sessions to provide feedback, ratings, and skill-gap analysis for clinicians.