Chad McDonald, Research Associate Professor in the College of Social Work 

Matt Davis, Research Associate Professor in thePopulation Health Sciences and Psychiatry 

The Role of AI in Content Development 

One of the most impressive aspects of this project is the integration of AI for content development. Generative AI is used to create nuanced scenarios that social workers might encounter in their professional interactions. This approach significantly reduces the time required for content creation, allowing for the generation of thousands of realistic and varied practice interactions. 

A Training Platform Inspired by Duolingo 

The training platform itself is reminiscent of Duolingo, but instead of teaching languages, it focuses on motivational interviewing. In addition to the chatbot, VMI currently includes hours of scaffolded exercises that range from simple to complex. This structure ensures that learners can build their skills progressively, starting with basic exercises and moving towards more challenging scenarios. 

Interactive Chatbot for Skill Practice 

A key component of the training platform is a chatbot developed using OpenAI's custom API. This chatbot acts as a client or caregiver, enabling social workers to practice their skills in a safe and effective learning environment. The chatbot provides feedback based on the learner’s interactions by switching from client mode to expert mode when prompted to grade the interaction. This feedback is crucial for helping professionals refine their use of motivational interviewing (MI) skills. 

Realistic Content Creation 

The AI's ability to generate realistic exercises and scenarios is a game-changer for professional training. For example, a social worker might interact with a chatbot character named Maria, who is involved in child welfare. The social worker selects goals and practices skills such as open-ended questions. The chatbot then provides feedback on the interaction, highlighting strengths and areas for improvement. This realistic practice is invaluable for preparing social workers for real-life situations. 

Proven Efficacy 

Chad and Matt explained the VMI app is rated very positively by students in terms of usability and relevance to social work practice. Additionally, they highlighted a randomized controlled trial that accessed the initially efficacy of the app compared to role plays; which found significant improvements pre/post at a rate no different than traditional lecture and role plays 

Conclusion 

The Virtual Skills & Workforce Trainer (VSWT) represents a significant advancement not only in social work training, but for many other behavioral health and medical professionals alike. By leveraging AI and virtual reality, Chad McDonald and Matt Davis have created a scalable and effective tool that provides personalized feedback and realistic practice opportunities. This innovative approach is poised to make a substantial impact on the health and human service field and beyond, enhancing the skills and preparedness of professionals across the board. 

For more information, you can reach out to Chad McDonald at  chad.mcdonald@utah.edu  or Matt Davis at  matt.davis@hsc.utah.edu. 

---

AI Community of Practice

This blog post is part of the ongoing work of the AI Community of Practice (ACP), a campus-wide group dedicated to fostering collaboration, knowledge sharing, and discussion around the use of artificial intelligence. The ACP provides a space for faculty, students, and staff to connect, exchange ideas, and explore emerging topics in AI.

The group meets on the fourth Tuesday of each month at 11:00 a.m. via Zoom.  All members of the campus community are welcome to attend and join the conversation.