Computational Models of Human Intuition

This article from Artificial Intuition Research (AIR) examines three approaches to computationally modelling aspects of human intuition: recognition-based decision-making, the discovery of non-obvious connections through semantic networks, and intuitive response patterns in large language models. Drawing on published research, it explores how these approaches represent different aspects of intuition and what their findings demonstrate. The article highlights an important distinction between computational systems producing outcomes that resemble human intuitive judgments and evidence that they use similar underlying cognitive processes. It discusses the limitations of the studies reviewed and identifies questions for further investigation into computational models of intuition. Developed from exploratory research undertaken during AIR’s Founding Research Internship Cohort, this article has been editorially reviewed and refined for publication. It has not undergone formal journal peer review.