Related Experiment Video
Updated: Aug 6, 2026
07:14
Virtual Agent for Real-Time Motivational Interviewing by Integrating Adaptive Nonverbal Behavior and Language Models
Published on: December 23, 2025
Can Motivational Interviewing be delivered using Artificial Intelligence chatbots? Evaluating the capability of
Candice Oster1, Richard Leibbrandt2, Judith Hocking1
1Flinders University, College of Nursing & Health Sciences, Caring Futures Institute, Adelaide, Australia.
Patient Education and Counseling
|July 23, 2026
Summary
Generative Pre-Trained Transformer (GPT-4o) shows promise in delivering Motivational Interviewing (MI) behaviors but needs improvement for proficiency. Further research is essential to enhance AI capabilities for accessible and cost-effective MI delivery.
Area of Science:
- Artificial Intelligence in Healthcare
- Behavioral Science
- Digital Health Interventions
Background:
- Motivational Interviewing (MI) is a patient-centered counseling approach to facilitate behavior change.
- Large Language Models (LLMs) offer potential for scalable digital health solutions.
- Evaluating AI's capability in delivering complex therapeutic techniques like MI is crucial.
Purpose of the Study:
- To systematically assess the proficiency of Generative Pre-Trained Transformer (GPT-4o) in delivering Motivational Interviewing.
- To analyze GPT-4o's adherence to MI principles and identify patterns in its conversational responses.
- To determine the potential of AI in enhancing accessibility and reducing costs of MI delivery.
Main Methods:
- OpenAI's GPT-4o was prompted to conduct simulated Motivational Interviews with patient actors.
- Interview transcripts were analyzed using the Motivational Interviewing Treatment Integrity (MITI) code.
- Conversation sequences were examined for patterned responses related to MI behaviors.
Main Results:
- GPT-4o exceeded 'good' competency thresholds for specific MI behaviors like complex reflections and reflection-to-question ratio.
- High counts of MI-adherent behaviors were observed, with minimal non-adherent behaviors.
- GPT-4o scored below 'fair' competency for global relational and technical skills, showing a tendency to shift towards persuasion.
Conclusions:
- GPT-4o demonstrates a capacity to mimic MI behaviors but requires further development for proficient delivery.
- AI-driven MI holds potential for increased accessibility and reduced costs.
- Further research into AI augmentation techniques is necessary before widespread implementation of LLMs like GPT-4o for MI.
Related Concept Videos
01:16Non-equilibrium in the Cell
An important concept in studying metabolism and energy is that of chemical equilibrium. Most chemical reactions are reversible. They can proceed in both directions, releasing energy into their environment in one direction, and absorbing it from the environment in the other direction. The same is true for the chemical reactions involved in cell metabolism, such as the breaking down and building up of proteins into and from individual amino acids, respectively. Reactants within a closed system...
01:25Motivational Bias
Cognitive bias results from limitations in thinking and information processing, leading to systematic errors in judgment. Conversely, motivational bias stems from personal desires or emotions, causing distortions in perception to align with self-interest. Motivational bias influences how individuals perceive and attribute causes to events, often shaped by personal needs, goals, and self-esteem preservation. This bias can distort judgment, leading to inaccurate assessments of success, failure,...