Guidelines for Debriefing AI Negotiation Exercises to Facilitate Learning

by Jeanne Brett and Virginia Stewart

It is very important that the instructor treats the AI negotiation exercise as seriously as the human-human negotiation exercises.

Make sure the students are comfortable using both the iDG platform and the AI negotiation platform. Even if your students are familiar with the iDG platform walk them through a simple AI negotiation exercise, such as Split $2 with… so that they can see how the AI platform works.

Make time for a normal exercise introduction, to prepare, to negotiate, to read and interact with the personalized feedback. Prepare a class-wide debrief that covers both the learning points of the exercise AND the experience negotiating with AI.  Assign a reflection on the exercise. You can use any of the suggested AI debriefing questions below.  Alternatively ask students to reflect, What did you learn about negotiating and about yourself as a negotiator from negotiating EXERCISE NAME with AI?

Suggested AI Debriefing Discussion Questions

1. How did you feel about negotiating with AI rather than a classmate? Why? You are likely to get a wide variety of answers to follow up both with the student answering and by asking other students Did you also feel that way, or did you feel differently?
   • Challenged – I knew I was negotiating with a counterpart that was prompted by an expert negotiator.
   • Confident/ Safe– I knew I could experiment and try new strategies without taking any reputational risks. I was relieved, I find my classmates verbally aggressive sometimes.
   • Frustrated – I don’t think a real person would negotiate like that. What did the AI do that made you think that? I had to do all the work – the AI wouldn’t share information/make counter-offers. The AI ended the negotiation without warning. I tried to expand the issues for negotiation, but the AI was not having it. My emotional approach was not reciprocated. Why do you suppose it wouldn’t? In retrospect is there something you might have done differently to get the AI to participate? Who else had this problem? How did you overcome it?
2. You can also ask this question about feelings before starting the AI exercise, during the exercise, after completing the exercise but before receiving feedback, or after receiving performance feedback. Expect and explore prenegotiation feelings: confident, curious, excited, calm, at ease, happy, nervous, apprehensive, unsure, cautious and hesitant.
    
   Or, you can also ask this question about feelings about the negotiation process and the negotiation outcome. [1] Expect post negotiation feelings: defeated, angry, frustrated, annoyed, overwhelmed, nervous, disappointed, dissatisfied, successful, happy, satisfied, more confident.
3. How did you treat the AI negotiator? Like you would treat a classmate in the role, or as a non-human, non-feeling entity? Why?  Answers vary, some students approach negotiating with AI as an opportunity to try out all the demanding, even unethical tactics that they would not use with a classmate. If the AI was properly prompted, it will not respond to such tactics and unless these students change their strategic approach, they are unlikely to reach agreement. This is actually a very good learning experience for students. You might want to bring into the discussion results from MIT bot competition that indicate higher performance outcomes associated with exhibitions of “warmth” (Vacarro et al. 2025). Remind students that AI models have been trained on books written for and by humans.
4. Do you think you learned more, as much, or less about negotiating when paired with an AI counterpart compared to a classmate? Why?  The answer to this question usually depends on how satisfied the student is with the outcome. If they are satisfied with the outcome, they are likely to think they learned a lot with the AI counterpart, particularly if it was challenging to reach an agreement. If appropriate discuss peer behaviors that support/inhibit classroom learning.
5. What did you learn about yourself negotiating with an AI counterpart?  Answers here are likely to vary but for the novice negotiator they are likely include: desire to be less hesitant, be warmer, regulate my emotions, be more persuasive, be more demanding.
6. Did you try to deceive the AI negotiator? Why? Do you think the AI negotiator was deceptive? This is an opportunity to discuss trustworthiness. Assessments of trustworthiness rely on perceptions of benevolence, integrity, and belief your counterpart cares about your well being. Why did you trust/distrust the AI? Is this trust misplaced?
7. What did you like about negotiating with AI as opposed to a classmate? Dislike? Given the propensity to negotiate is quite low in the general population with many people being willing to pay hundreds if not thousands to avoid negotiating (Hunsaker, Zhang, & Lee, 2025), what role do you see for deploying AI negotiators in your chosen career? This is an opportunity to bring in the range of roles for AI in negotiations, from AI-assisted, to semi-autonomous, to fully autonomous and provide field examples of each (Revilla & Saenz, 2025).
8. What did you learn from the personalized feedback? You can discuss the quality of the feedback including accuracy, speed, and motivational impact. If there is time, the class can explore the value of AI feedback provided in other work and life situations and when it is impactful or dismissed. Is positive feedback from an AI on par with that from a peer or instructor? Is it preferable to receive negative feedback from a trusted mentor or from an AI?
   
   
   
   [1] Stewart and Dowling-Heatherington report that students dissatisfied with their outcome also tend to be dissatisfied with their personalized feedback. Rating the AI as fair and honest seems to be critical to maximizing the learning from the exercise.

Notes: These Guidelines are based on Jeanne Brett’s assessment of student learning from a review of empirical research, AI and Negotiation Learning submitted to IACM 2026, but even more on Virginia Stewart and Linda Dowling-Hetherington’s report on their NTR grant, Learning to Negotiate in the Age of AI also submitted to IACM 2026. Please see these documents for references.