🌍 Organizing a geopolitical role-playing game in geography with AI: method

The role-playing game on an actor conflict is an old and well-documented format in geography — Géoconfluences (ENS Lyon) has been publishing implementations for years, and Model United Nations (MUN) type simulations are practiced in several academies in HGGSP. What conversational AI changes is not the principle — having students embody opposing viewpoints — but the preparation: each actor can become a character to question directly, rather than a paper sheet to read silently before playing a role they poorly master.

Two well-established formats, one same principle

The 'actor conflict' is a full-fledged geography concept (see the Géoconfluences glossary): a conflict opposing groups of actors on a territory, typically a development project where a developmentalist camp (growth, jobs) opposes an environmentalist camp (nature protection). Géoconfluences documents, for example, a classroom implementation around the 'shark crisis' in Réunion, with the issue: how to manage the shark risk sustainably? Students receive documents to prepare their role, then confront their positions.

The Model United Nations type simulation follows a similar but more formalized logic: the class is divided into delegations (often pairs), each delegation represents a country on a negotiation theme, and the official format includes documentary research, drafting a position paper, then a structured debate according to international body rules. The United Nations itself publishes an official guide to organize it, and several academies (Nantes, Aix-Marseille, Toulouse) have documented concrete implementations in high school.

Step 1: choose a specific actor conflict, anchored on a territory

As with an educational escape game, a broad theme ('globalization', 'sustainable development') offers nothing to play: no one knows what to defend. An actor conflict needs a specific territory and a concrete issue on which real interests oppose — a local development project, a disputed resource, a risk to manage. The more precise the issue ('Should such a project be authorized on such a territory?' rather than 'Territorial planning'), the more defensible and distinct the actors' positions become.

Step 2: provide an actor to question, not just a sheet to read

The classic format hands each student a role sheet, who must then improvise a position they rarely master well on the first try. With a conversational AI character on Parlova, the student can first question the actor they will then embody — a mayor in favor of the project, a local residents' association opposed, an economic operator of the territory — before playing the role themselves in class. This preparation step through dialogue changes the nature of the subsequent improvisation: the student arrives with arguments they have obtained themselves, not just read.

Step 3: adjust the format according to the available time

A simple actor conflict fits into a one-hour session: ten minutes of territory and issue presentation, twenty minutes of actor questioning independently (via Parlova or documents), twenty minutes of oral confrontation in class, ten minutes of collective decision. A complete UN simulation requires more — the classic academic format spans several sessions (institution presentation, country research, position drafting, final debate) — but can be compressed into a single long session if documentary research is replaced by dialogue with the country-character.

Step 4: build a decision, not just a simple debate

A debate that ends without a decision leaves the impression of an opinion exercise rather than a geographical exercise. The session must result in a concrete production: an argued development recommendation (case of actor conflict), or a resolution voted according to a format close to that of real bodies (case of UN simulation). It's not the 'right answer' that matters — several different recommendations can be equally defensible — but the ability to justify an arbitration based on positions actually gathered during preparation.

Step 5: prepare the debrief before the session

As with any immersive format, the debrief is prepared in advance, not improvised. Decide in advance the program concepts you want to highlight — spatial actors, sustainable development, territory governance, conflict scales — and prepare two or three questions that force students to step out of the role they just played to analyze the situation with the geographer's tools. Without this time for reflection, the session remains a successful role-playing game but not necessarily a geography lesson retained.

What this format does not replace

Questioning an AI actor before playing their role helps to appropriate a position, but does not replace working on real primary sources (local press articles, numerical data, maps) that give the conflict its true documentary depth — Géoconfluences bases its own implementations on a scientific article distributed to students, not on an invented sheet. The AI prepares the role embodiment; the documentary file remains what anchors the exercise in a real geographical case rather than in a plausible fiction.

Frequently asked questions

1. 1How many actors are needed for an actor conflict in one hour? Three are generally sufficient: two opposing positions and an ambivalent actor that avoids the caricature of two blocks.
2. 2Is it necessary to vote on a UN-style resolution? No for a simple actor conflict — an argued recommendation by group is enough. The formal vote is mainly useful for a real, longer, and more institutional UN simulation.
3. 3Is this format also suitable for middle school? The simple actor conflict (a local development project, three positions) works from 4e-3e; the complete UN simulation, with position paper and debate procedure, is better suited for high school (notably HGGSP).