ChatGPT and the Technical Ontologies of Critical Political Science

Abstract The following paper asks what forms of critical political science are possible given the advent of Machine Learning algorithms like ChatGPT? The technical conditions that provoke this question stem from the fact that such automated systems do not generate identities and representations (as do traditional media with which critical political science conventionally contends) but produce instead rendered outputs based on Bayesian probability calculations. Thus the challenge before us is to theorize criticism vis-a-vis the technical ontologies of the algorithm dispositif.