"

8 Policy Advocacy Roleplay with AI

Karen Magruder, LCSW-S

Author Bio

Karen Magruder, LCSW-S is an Associate Professor of Practice at the University of Texas at Arlington School of Social Work, where she brings a broad background in aging, mental health, and environmental justice to the classroom. A dedicated and award-winning educator, Karen is passionate about integrating evidence-based and innovative techniques to promote student success in online learning. Karen is also pursuing her Doctor of Social Work (DSW) degree at the University of Kentucky, where she is studying the implications of AI for social work practice and education. She manages a small private therapy practice, provides clinical supervision, and maintains a free social work education resources YouTube channel.

Course Context

This activity is designed to be adaptable to a variety of contexts, but it was originally designed with the following student population in mind:

  • Discipline: Social Work
  • Level: Undergraduate
  • Course Name: Social Welfare Policy & Services
  • Modality: In-person
  • Context: Survey of social welfare services, programs, and policies, including those that impact children & families, older adults, healthcare, and mental health. Emphasis on advocacy for macro-level policies that support client wellbeing and social justice.
  • Number of Students: 25

Activity Overview

This experiential learning activity utilizes ChatGPT, or an equivalent generative AI application, to simulate an interaction with an elected representative. Students engage in roleplay, with the AI acting as a staffer communicating with a social worker via email, allowing them to gain confidence with advocacy skills in a controlled and low-stakes environment. This approach provides a safe space to develop persuasion and communication skills before transitioning to real-world interactions. By leveraging AI, the simulation offers an endless variety of possible political stances and reactions, adapts to student input, and provides personalized feedback for enhanced learning. This activity can be adapted to other disciplines that engage in client interactions, such as a customer service chat, handling sales inquiries, nursing telehealth portal, or a teacher messaging a parent.

Directions

For Instructors

  • Prepare students for this activity by asking them to be ready to discuss a specific topic as a class, likely related to that day’s learning outcomes. For instance, you might assign them to research a particular social welfare policy, such as affordable housing initiatives, child welfare reforms, or affordable mental health services. Encourage them to gather relevant data, understand different viewpoints, and develop clear arguments for why the policy should be supported. This preparation will enable them to engage more effectively in the roleplay and apply their learning in a practical context.
  • Prior to class, create/log in to your AI account, such as ChatGPT. This activity was originally implemented with ChatGPT’s free version (3.5) but may be replicated using equivalent generative AI chat-based platforms.
  • Prime your selected AI tool about the purpose of the activity, its role in the scenario, and any other parameters you’d like. For example, “In this chat thread, we’ll be doing a roleplay where a social work student is lobbying for a social welfare policy, and you are the staffer of an elected official. I will start with the phrase ‘Beginning now’ to signal the start and type ‘End roleplay’ when finished.” You may also provide any additional guidance, such as preferred political stance, attitude, receptiveness, etc. For instance, you might prompt, “…Please adopt a neutral but respectful and professional tone, ask detailed questions about the policy proposal, and be moderately receptive but critical to ensure thorough discussion. Your current budget is very tight, and the majority of your constituents support [xyz]. Also be sure to ask the social worker what their qualifications are for raising these suggestions.”
  • Explain the purpose and logistics of the activity (see Instructions for Students)
  • Share your computer screen, projected to a screen at the front of the class (or a screen share if online), and model the first interaction (see Instructions for Students).
  • As the activity progresses, you may wish to provide real-time guidance or seek input from the class. For example, if students appear to be struggling with formulating their arguments or need help responding to questions posed by the AI, you can pause the roleplay and offer suggestions or examples. You might say, “Let’s pause for a moment. What are some key points we want to emphasize in our advocacy?” or “How might we address the staffer’s concerns about funding for this policy?” This can help students stay engaged and ensure they are effectively learning from the exercise.
  • After the activity has concluded, ask ChatGPT to provide feedback about the interaction. You may ask for general impressions or specify a particular skill, such as supporting recommendations with data.
  • Debrief with the class, discussing any particularly persuasive techniques, missed opportunities, lessons learned, etc.

For Students

In this activity, we are going to work together as a class to play the part of a social worker who is advocating for [pre-selected policy, which they are prepared for] with a political office staffer [specify specific office, such as city council member, governor, mayor]. The “staffer”, an assistant who works with the elected official will be simulated by an artificial intelligence app called ChatGPT. Our goal here is to [insert learning outcome or focus area].

Here’s how it will work. We’ll spend about [x] minutes engaging in a dialogue with this staffer. Every student will have the opportunity to contribute a response. Remember that this is a low-stakes activity because it’s not a real political office, and so there’s no “wrong answer” or feeling foolish, just learning opportunities. We are here to support and learn from each other. The way it will work is [Instructor chooses one approach]:

Option 1: Each time the staffer responds with a comment or question, raise your hand if you’d like to be the responding social worker.

Option 2: One by one, I will call on students to give a response as the social worker. If you are stuck, you may “phone a friend” from the class to assist you. [If the dialogue reaches saturation before all students have a turn, start a new dialogue with a “new staffer” or about a different policy]

Option 3: Each time the staffer responds to us, we will discuss as a class how we might respond and come up with a response together.

I have already set up the roleplay. Our first step will be to initiate the conversation by introducing ourselves and stating the policy we are advocating for. [Instructor demonstrates first interactions with “staffer.” resulting in a response from ChatGPT]. Okay, the staffer has responded [read response out loud]. What should we say next? [continue until the conclusion of roleplay]

Benefit to Students

After completing this exercise, students will be able to demonstrate:

  • Professionalism in written communication
  • Critical thinking and problem-solving abilities
  • Perseverance in advocacy despite resistance

Assessment

This activity was originally designed to be an ungraded in-class learning experience and thus does not include an assessment component. However, instructors can modify this activity into a graded assignment by asking students to submit individual chat logs, along with a reflection on the experience and application to professional practice.

Cross-Disciplinary Applications

There are a number of disciplines which could adopt this activity to simulate interactions with clients or customers, including the following:

Business and Management: Customer Service and Negotiation Scenarios

The GenAI tool can simulate customer interactions where students role-play as customer service representatives or salespeople. They can practice handling complaints, negotiating deals, and providing product information.

Healthcare: Patient Consultations

The AI can act as a patient presenting various symptoms or health concerns, allowing medical or nursing students to practice patient interviews, diagnostic questioning, and bedside manner.

Education: Parent-Teacher Conferences

An AI app can play the role of parents with different concerns about their child’s performance or behavior, enabling education students to practice conducting parent-teacher conferences.

Disclosures

AI-assisted content. Mavs Open Press defines work as AI-assisted when author-created content is enhanced, organized, or edited using an AI tool. 

This OER was created using AI assistance. 

License

Icon for the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License

AI-Powered Education: Innovative Teaching Strategies to Elevate Student Learning Copyright © 2025 by Karen Magruder, LCSW-S; Ann M.L. Cavallo, PhD; Andrew M. Clark, PhD; Karen L. Bravo, PhD, MSN, RN; Jess Kahlow, PhD; Christy Spivey, PhD; Heather E. Philip, PhD; Kevin Carr, PhD; Michael Buckman, MBA; Jennifer Roye, EdD, MSN, RN, CHSE-A, CNE; Hugh J.D. Kellam, PhD; Luis E. Pérez Cortés, PhD; and Rosie Kallie, PhD is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, except where otherwise noted.