Mapping the jagged edges of AI with the tools of cognitive science

Current artificial intelligence systems demonstrate a surprising amount of heterogeneity in their abilities, displaying superhuman competence in some tasks but puzzling limitations in others. In this talk, Griffiths will argue that the tools we need for understanding this heterogeneity can be found in cognitive science, where researchers have spent decades developing theoretical and empirical methods for making sense of the capabilities of intelligent systems. Griffiths will highlight ways in which empirical methods from cognitive science—tools for studying similarity, categorization, and reasoning—can shed light on the capabilities of AI systems.