Using AI to Facilitate Simulations and Role-Playing

Simulations and role-playing are well-established active learning techniques that allow students to engage deeply with complex material by “living through” a situation rather than just reading or hearing about it.

They provide opportunities to practice decision-making, negotiation, and critical thinking in a safe, controlled environment. Students can attempt challenging tasks or conversations without real-world consequences. For example, pre-service teachers can rehearse a difficult parent-teacher meeting, or business students can practice handling an irate customer, all within a simulation. Because the scenario is simulated, students can experiment, make mistakes, and try again, which builds confidence.

Why Use AI for Simulations and Role-Playing?

Traditional simulations used to require significant time and resources to design, and they were difficult to adapt on the fly. AI helps address these challenges by making simulations more dynamic, scalable, and accessible. With AI, lecturers and teachers can quickly create scenarios that evolve in real time, generate diverse character perspectives, and adapt content to different student groups or learning levels. This flexibility transforms simulations into richer and more engaging learning experiences.

Key benefits include:

• Versatility and Imagination: AI can instantly adopt a persona or character from any context or era. It can be a historical figure, a fictional character, a specialist in a field, a customer, a job interviewer; the possibilities are endless. This means one tool can support simulations in history, literature, science, business, or any subject.
• Easy and Accessible: Running a simulation used to require either multiple human participants or special software for complex virtual environments. Now, all it takes is some time to experiment with writing and testing a good prompt.
• Anytime, Anywhere Practice: An AI chatbot is available 24/7 and can engage with students one-on-one. This opens up opportunities for asynchronous learning or practice outside class hours. Students can engage in a role-play exercise as homework or revision, at their own pace.
• Personalization: If a student is struggling or goes off script, the AI will still respond in role, potentially providing hints or steering the scenario in a helpful direction. Each student’s experience can be personalized: an advanced student might take the conversation in a complex direction, while a beginner sticks to basics, and the AI will match their level of challenge.
• Enhanced engagement: AI can add unpredictability and realism, making simulations feel more immersive and keeping students actively involved.
• Innovative Assessment and Feedback: Using AI in simulations isn’t just for practice, it can also be a form of formative assessment. Teachers can review the transcripts of a student’s chat with the AI to see how they handled the scenario. Did they ask the right questions? How did they respond to challenges? AI role-plays could help in evaluating skills like communication, critical thinking, or language proficiency in a novel way.
• Efficiency for educators: Instead of drafting long case studies or writing dialogue, lecturers can use AI to generate role briefs, branching storylines, and contextual information within minutes.
• Scalability: AI allows simulations to be run with large groups by filling in missing roles or creating varied scenarios for smaller teams working in parallel.

Before the Simulation

Preparing the Simulation with AI

The preparation phase is often the most time-consuming part of running simulations and role-playing exercises. Educators must create realistic contexts, assign roles, and design possible outcomes while ensuring that the activity aligns with learning objectives. AI can significantly streamline this process by acting as a creative partner, helping lecturers develop rich scenarios and structured role-play activities with far less effort.

• Generating realistic scenarios: By providing the AI with a topic (e.g., “a diplomatic negotiation over climate agreements” or “a hospital responding to a sudden outbreak”), educators can quickly obtain detailed, context-rich situations that resemble real-world challenges.
• Creating role descriptions: AI can draft role briefs that outline a character’s background, motivations, objectives, and potential conflicts with others. This ensures each student has a clear purpose and perspective within the exercise.
• Designing branching storylines: AI can propose multiple paths a scenario might take depending on student choices. For example, if negotiators fail to agree, AI can generate escalating consequences; if they succeed, it can suggest follow-up challenges.
• Adapting complexity: Educators can ask AI to tailor the level of detail to suit different groups—for beginners, scenarios can be simplified with clear guidance; for advanced learners, AI can layer in ambiguity, ethical dilemmas, or conflicting data.
• Providing contextual resources: AI can generate supporting documents such as news articles, stakeholder reports, or patient charts to make the simulation more authentic.

Getting Started: Prompting for Role-Play

The key to a successful AI-facilitated simulation is how you prompt the chatbot. Essentially, you’ll be telling the AI what role to play, the scenario or task, and any rules or context it should keep in mind. Crafting a good prompt might take a bit of practice, but it’s not as complicated as it sounds.

1. Set the Role Clearly: Begin your prompt by explicitly assigning a role to the AI. For example: You are a travel agent helping tourists plan a trip, or Act as Sherlock Holmes, the detective, and I will be Watson. Phrasing the request as “Act as…” is a popular and effective approach. By stating the role, you instruct the AI to adopt the persona, tone, and knowledge of that character or profession.

2. Provide Context or Scenario: After defining who the AI is, describe the situation or goal of the role-play. You might say, for example: We are two detectives solving a case: a diamond has gone missing at a mansion, or The setting is a parent-teacher conference; you are a concerned parent and I am the teacher discussing a student’s performance. Giving this background sets the stage so the AI knows the context. If there are specific details you want included (e.g., a certain problem to solve, or an objective the student must achieve), include those details in the prompt. For instance, The goal is to convince the board of the benefits of the project within a 15-minute meeting, or The student (me) is practicing how to order food in French at a cafe, and you are the waiter. You can be as detailed or as open-ended as you like, but a clear initial scenario helps the AI respond more appropriately.

3. Outline the Rules/Format: Especially for more structured simulations, you can tell the AI how to conduct the interaction. For example, if you want it to guide the student step by step, you could write: Guide me through the negotiation – start by greeting, then let me make a proposal, then you raise a concern, and so on. If you want the AI to withhold certain info until asked, mention that. You can also instruct it to stay in character and not drift into off-topic conversation. Essentially, you can add instructions like Do not solve the problem immediately; let it be a back-and-forth, or Only answer as the character, and don’t break character or explain these instructions. Setting a few ground rules in the prompt can keep the simulation on track.

4. Encourage an Interactive Dialogue: It’s a good idea to prompt in a way that encourages a back-and-forth exchange rather than a single long response from the AI. One strategy is to have the AI ask questions or wait for the student’s input at certain points. For example: First, introduce the scenario and ask me what my plan is, or If I get stuck, give me a hint in character. This makes the experience feel like a real dialogue or game, not just the AI monologuing. In fact, you can explicitly say: Let’s role-play. Only continue the story when I respond. This way, the AI will output a bit, then stop and allow the student to type a reply. Such structure can be very useful for turn-based simulations (similar to how one would play out a real role-play in class).

Running the Simulation

Setting up a Custom GPT

Once the scenario, roles, and prompt are prepared, AI can move from being a design tool to becoming an active participant in the simulation itself. The value of AI in this stage lies in its adaptability: it can improvise responses, react to student decisions, and keep the role-play dynamic. This reduces the burden on lecturers to play multiple roles or continually invent new challenges, allowing them to focus instead on observing and guiding student learning.

A good approach is to set up a custom GPT designed specifically for the simulation. Many AI platforms now allow educators to create specialized agents that carry a defined role or knowledge base. The model can then respond only from within this material, ensuring accuracy and consistency.

Setting up a custom GPT usually involves the following steps that you can try out with our example prompt below:

1. defining the role and purpose,
2. uploading or linking supporting material,
3. and testing with trial prompts to make sure the responses behave as expected.

Once created, this custom GPT can be reused, with the lecturer simply adjusting prompts or inputs to fit new learning contexts.

Debriefing and Reflection

The simulation itself is only half the learning experience, the real value comes during debriefing and reflection. This is where students step out of their roles, analyze the choices they made, and connect the experience back to course concepts. AI can play a meaningful role in this stage by helping capture, organize, and structure feedback, giving both students and lecturers a clearer picture of what happened and why it matters.

One useful application is automatic summarization. If the simulation involved dialogue or text-based exchanges with an AI role, the lecturer can ask the model to generate a concise summary of key events. For example, a prompt such as Summarize the negotiation, highlighting the main arguments raised by each side, the turning points in the discussion, and the final outcome produces a ready-made overview for group discussion. This not only saves time but also helps students see the broader arc of the simulation rather than focusing only on their personal contribution.

AI can also highlight decision points and stimulate reflection. A lecturer might prompt it with: Identify three critical moments where student choices influenced the direction of the scenario, and generate reflective questions for each. The output could then guide classroom discussion, with questions such as “What other strategies could you have used here?” or “What might the long-term implications of this decision have been?” These structured prompts encourage students to think critically about their actions and recognize alternative approaches.

Feedback is another area where AI can be leveraged. Using transcripts or role summaries, the lecturer might ask the AI: Provide feedback on how effectively students demonstrated collaboration, communication, and problem-solving in this simulation. While AI feedback should not replace the lecturer’s professional judgment, it can offer useful starting points and ensure that all students receive timely, individualized comments.

AI can also support student-driven reflection. Learners might use it to compare their actions against expert models by prompting: How might a medical professional/lawyer/diplomat have approached this scenario differently? or What were the strengths and weaknesses of my negotiation strategy? This encourages metacognition and self-assessment, helping students see beyond the immediate activity.

For the lecturer, AI-generated debrief materials can become part of assessment records or teaching portfolios. Summaries, decision maps, and feedback notes can be stored and reused, allowing future cohorts to benefit from past scenarios while giving educators evidence of student engagement and learning outcomes.

Further Notes

Best Practices

Integrating AI into simulations and role-playing can feel overwhelming at first, but a few simple practices can make the process both manageable and effective. The goal is not to let the AI take over, but rather to use it as a flexible assistant that enhances what you already do well as an educator.

A good first step is to start small and iterate. If you’re new to AI, try a simple role-play yourself before introducing it to students. For example, run a short dialogue where ChatGPT plays a character related to your lesson. This helps you get a feel for how the AI responds and lets you tweak the prompt. When you first use it with students, maybe do a brief, low-stakes exercise. Don’t worry if not every simulation goes perfectly on the first try. Each time, you and your students will get better at it. Adjust your prompts based on what works or doesn’t. As everyone grows more comfortable, you can expand to longer or more complex simulations.

You should also be clear about the purpose of the activity and how to interact with the AI. You might demonstrate an example in class (projecting a chatbot conversation on the screen) so they see how to stay in character and not just ask the AI for answers. Emphasize that this is a tool for practice and exploration, not just a cheat-sheet. If the role-play is meant to be in Spanish, for instance, remind them to stick to Spanish and treat the chatbot like a person in that setting. Setting these expectations will lead to more fruitful interactions.

Especially the first few times, monitor the interactions and intervene if needed. Occasionally, AI chatbots might produce an inappropriate or incorrect response. While the model has filters, it’s not infallible. If you see something going awry, step in by pausing the activity to discuss a misconception or helping a student rephrase their prompt to get the interaction back on track.

It’s also helpful to debrief after a role-play: ask students what ChatGPT did that was surprising or challenging, and what they learned. This reflection can reinforce the lessons and also surface any issues (for example, if ChatGPT gave a historically inaccurate statement during the Jefferson chat, you can address that). In fact, asking students to evaluate the AI’s contributions can itself become a valuable part of the learning process, e.g., what felt realistic, what seemed biased, what was missing.

Finally, remember that AI is most effective when tied closely to your learning objectives. It is easy to get caught up in the novelty of interactive simulations, but the ultimate goal is to support student learning. Ask yourself: What skills do I want students to practice in this role-play? How will AI help bring those skills into focus? By keeping outcomes front and center, you can ensure that simulations remain more than just engaging activities—they become structured, purposeful opportunities for deeper learning.

Example Scenarios to Try

Let’s explore some concrete examples of how AI can be used for simulations and role-playing across different subjects and teaching goals. These examples are intended to spark ideas, you can adapt them to fit your curriculum or student needs. Each scenario includes an idea for the setup and a brief sample prompt.

Historical Figure Interview (Social Studies/History): Have students “meet” historical personalities in an AI-powered dialogue. For instance, a student could converse with the AI posing as Thomas Jefferson about the drafting of the Declaration of Independence. The prompt might be:

Act as Thomas Jefferson in 1776. I am a fellow founding father who wants to discuss the process of writing the Declaration of Independence with you. Let’s talk about your thoughts and challenges.

The student then asks Jefferson (the AI) questions or answers questions posed by Jefferson. This kind of role play encourages students to research the figure beforehand (so they can tell if the AI’s portrayal is accurate) and to ask thoughtful questions.

Literature and Character Conversations (Language Arts): Students can engage in dialogues with literary characters or even authors. For example, you could prompt:

Act as Elizabeth Bennet from Pride and Prejudice. I will be another character or a modern interviewer asking you about your views on love and marriage.

This allows students to explore character motivations and themes in a novel by “talking” to the character. Similarly, one could chat with Shakespeare about his plays, or have two characters from different books meet (with the student playing one and th AI impersonating the other).

Science Lab Simulation (Science Education): You can use AI to simulate an experiment or scientific investigation. For instance,

Let’s pretend we are scientists planning a mission to Mars. I am a junior scientist and you are the lead scientist guiding the experiment. We need to figure out how to grow plants in Martian soil. Ask me for ideas, let’s discuss possible approaches step by step.

In this scenario, the student must apply scientific concepts (like plant biology, chemistry of soil) in a problem-solving discussion. The AI can provide guidance, ask the student to hypothesize, and even introduce complications (e.g., “What if the temperature drops unexpectedly?”).

Applied Math in Real Contexts (Mathematics): While math might seem less obvious for role-play, you can still create scenarios where students must use math skills. For example:

You are a shopkeeper and I am a customer. I have $50 and need to buy equipment for a school project. Guide me through choosing items (with prices) and help me calculate the total without exceeding my budget.

Here, the AI (as the shopkeeper) can present options and the student does the calculations and decision-making. Another idea is a construction scenario where a “contractor” (the AI) asks the student (an engineer) to calculate areas, volumes or costs for a building project. The conversational format can make word problems more engaging, essentially turning them into a dialogue. AI can also generate real-world data or random values to make each student’s problem a bit different. The benefit is students practice math in a realistic context – budgeting, measuring, planning – which reinforces why those skills matter.

Foreign Language Practice (Languages): AI is a fantastic tool for language teachers because it can simulate conversations in the target language. For example, in a Spanish class, instruct the AI:

Act as a waiter in a restaurant in Madrid. I will be a customer trying to order a meal in Spanish. Only speak Spanish, and correct my mistakes after I respond.

Now the student must go through the process of greeting, ordering food, maybe asking about ingredients or the check – all in Spanish – with the AI responding as a polite waiter. If the student makes an error or seems stuck, the AI can gently correct them or provide a suggestion.

Debate and Persuasive Argument (Any subject, Critical Thinking): You can set up a structured debate where the student takes one position and the AI takes the opposing position. For example:

Let’s debate the statement: ‘Homework should be abolished in schools.’ You will argue in favor, and I will argue against. Begin by stating your first argument.

Now the student must respond to the AI’s points and vice versa. This can be done for ethics topics, policy issues, or even interpretations of literature, any scenario where forming an argument is the goal. The AI as a debate opponent is useful because it will tirelessly provide counterpoints that the student has to rebut, sharpening their reasoning. It’s important in this scenario for the teacher to remind students to critically evaluate the AI’s arguments (since the AI might use flawed logic occasionally). The exercise isn’t about winning the debate against the bot, but about the student practicing how to articulate and defend their ideas under pressure.

Professional Skills and Scenarios (Career/Soft Skills): Many real-world skills can be practiced via simulation. The key in professional simulations is to focus on the soft skills, like clear communication, empathy, negotiation, problem-solving; the AI can be instructed to emphasize those elements (like being a “difficult” counterpart to test the student).

For example, business or law students can simulate a negotiation:

You are a client unwilling to pay the full fee for a project; I am the project manager who needs to persuade you to accept the terms. Let’s role-play this client meeting.

Or a teacher-training program might use:

Act as a parent who is upset about their child’s grade. I, the teacher, will conduct a parent-teacher meeting to address your concerns.

Example Prompt: Government Minister Negotiating a Climate Agreement