Lay Beliefs About Artificial Versus Artificial Intelligence: Rethinking Theory of Machine

Research on augmented judgment and decision-making—where users retain responsibility for the final decision but receive input from algorithms prior to or during the judgment process—has largely contrasted human and algorithmic sources of judgment. Accordingly, Logg’s (2022) “Theory of Machine” is a conceptual framework for describing and analyzing people’s lay theories about how human and algorithmic judgment differ. Indeed, people often treat humans and algorithms as different kinds, that is, functionally distinct ontological entities. Therefore, I propose to complement the predominant human-centric lens on algorithmic judgment with an explicit algorithm-centric one, focusing on people’s lay theories about how different algorithms differ. Put differently, the core psychological claim of Theory of Machine 2.0 is that lay perceivers also differentiate among various AI systems. The first main contribution is the synthesis of capability contrasts across AI systems. People’s perceptions of these contrasts may be formed and shaped by personal experience and media exposure. Most importantly, their expectations and beliefs are supposed to consequentially guide their downstream user behavior—such as system trust, algorithmic advice weighting, and accountability attribution. The second main contribution is the proposal of testable research questions and designs for more algorithm-centric future research on people’s augmented judgment and decision-making. The theoretical perspective proposed in this article clarifies how people form and use lay theories about different AI systems and offers practical levers for the design, deployment, and evaluation of algorithmic decision-support systems.