Case study: AI & simulations

Heiko Pääbo (University of Tartu): How to enhance IR simulations with an AI tool?

The AI-assisted simulations were conducted as part of the Baltic Sea Region Studies specialisation module within the Master’s programme in International Relations and Regional Studies. The primary learning outcome of the assignment was for students to manage an escalating security crisis and reflect on their actions. Prior to the simulation, students acquired knowledge of the overall foreign policy objectives of the regional countries, their engagement in regional cooperation, and an overview of the security environment in the Baltic Sea region. This foundational understanding was crucial in preparing students to navigate the complexities of the simulated crisis effectively. The simulations, involving 20 students in 2023 and 14 in 2024, were conducted online using Moodle forums, providing a platform for interactive engagement and collaboration.

Generative AI (GenAI) tools, such as ChatGPT (paid version 4.0) and Microsoft Copilot, were employed in the simulations to enhance the learning experience. While both tools had their advantages and limitations, ChatGPT was utilised more frequently due to its ability to provide more nuanced responses. However, the variation in the tools also allowed for changes in the tone and style of the output, contributing to a more realistic portrayal of the crisis. GenAI supported the simulations by generating realistic content, adapting to student actions, and offering feedback. The AI-generated content created an immersive and dynamic environment, challenging students to think critically and adapt to evolving situations. Moreover, the AI’s ability to analyse student decisions and provide personalised feedback facilitated self-reflection and the identification of key learnings from the simulation.