Read the preprint on SSRN. Reviews are welcome.
Abstract: The pervasive influence of social media algorithms on public opinion and the formation of social realities necessitates transparent and accessible tools for exploration and analysis. This paper formally introduces the Digital Poiesis Laboratory, an interactive agent-based modeling (ABM) framework designed to demystify the mechanisms of online opinion formation and algorithmic influence. We detail the model’s core components-including agent psychology grounded in identity fusion theory, homophily-driven network structures, and a novel hybrid algorithmic feed model-presenting the underlying mathematical formulations. We also describe the laboratory’s open-source user interface, which serves as an exploratory instrument for policymakers, researchers, and the informed public. Through an illustrative use case comparing “Facebooklike” and “X-like” platform archetypes, we demonstrate the model’s utility in building mechanistic intuition about how algorithmic control shapes perception and opinion dynamics. We conclude by discussing the model’s current limitations and outlining a clear agenda for future research, including extensions for dynamic networks and empirical calibration.