The AIM-3D Project at Notre Dame team organized and hosted a panel session at the MPSA conference titled “Modeling Democracy with AI: Computational Approaches to Democratic Development and Decline.”
Michael Coppedge, Valentina (Valya) Kuskova, and Dmitry Zaytsev presented “AIM-3D: A Prototype Neural Network for Modeling Democratization,” showcasing early work on modeling and forecasting long-term democratic trajectories using AI.
We were also joined by an excellent lineup of presentations:
Andreas Beger and Rick Morgan – Forecasting Shifts in Democratic Governance: A Review of the Democratic Space Barometer Project – including reflections on earlier efforts to forecast democratization using machine learning in the V-Dem Forecast project.
Matthew Hauenstein – Patronage Power-Sharing and Rebel Group Splintering – highlighting how temporary power-sharing arrangements after civil wars can stabilize political coalitions, offering important insights for modeling how democratic institutions may emerge or endure.
Giovanni Pagano – Bending Bureaucracy: Detecting Political Capture of Administration with Computational Legislative Analysis – using large language models and computational analysis of legislative texts to trace how bureaucratic reforms may enable political capture under democratic backsliding.
We are especially grateful to our discussants, Brigitte Aimee Seim and Alexei Zakharov, for their insightful comments and engaging discussion.
A great exchange at the intersection of AI, computational methods and political science – looking forward to continuing these conversations and collaborations.
