Mapping the Computational Turn: Artificial Intelligence and Machine Learning in Migration Research

This study systematically maps the applications of artificial intelligence (AI) and machine learning (ML) in migration studies. We searched the academic literature published in major databases between 1990 and 2025, conducting an in-depth analysis of 117 search records that used computational methods to investigate migration-related research questions. The study disaggregated four AI approaches (Rationally Acting AI, Humanly Acting AI, Rationally Thinking AI, Humanly Thinking AI) and three ML approaches (Supervised ML, Unsupervised ML, Reinforcement ML) then intersected them with key research questions in migration studies. The study makes three main contributions. First, it reveals Supervised ML and Rationally Acting AI as the most preferred computational approaches in migration studies, clearly dominating over other computational approaches of study. Second, it transposes each computational approach to the research questions of study, identifying methodological patterns, disciplinary preferences, and opportunities for novel applications. Third, and most importantly, computational migration research continues to be concentrated on forecasting, prediction, optimization, and classification about migrants as opposed to computational research being conducted for migrants. Accordingly, we set out an agenda and call for further computational migration research and design centering migrant rights and wellbeing rather than solely supporting the institutional and administrative efficiency of migration management.