Individual and team profiling to support theory of mind in artificial social intelligence

Rhyse Bendell,Jessica Williams,Stephen M. Fiore,Florian Jentsch
DOI: https://doi.org/10.1038/s41598-024-63122-8
IF: 4.6
2024-06-03
Scientific Reports
Abstract:We describe an approach aimed at helping artificial intelligence develop theory of mind of their human teammates to support team interactions. We show how this can be supported through the provision of quantifiable, machine-readable, a priori information about the human team members to an agent. We first show how our profiling approach can capture individual team member characteristic profiles that can be constructed from sparse data and provided to agents to support the development of artificial theory of mind. We then show how it captures features of team composition that may influence team performance. We document this through an experiment examining factors influencing the performance of ad-hoc teams executing a complex team coordination task when paired with an artificial social intelligence (ASI) teammate. We report the relationship between the individual and team characteristics and measures related to task performance and self-reported perceptions of the ASI. The results show that individual and emergent team profiles were able to characterize features of the team that predicted behavior and explain differences in perceptions of ASI. Further, the features of these profiles may interact differently when teams work with human versus ASI advisors. Most strikingly, our analyses showed that ASI advisors had a strong positive impact on low potential teams such that they improved the performance of those teams across mission outcome measures. We discuss these findings in the context of developing intelligent technologies capable of social cognition and engage in collaborative behaviors that improve team effectiveness.
multidisciplinary sciences
What problem does this paper attempt to address?
The core problem that this paper attempts to solve is to develop a method that enables artificial intelligence to develop a "Theory of Mind" (ToM) to support team interactions with human teammates. Specifically, the paper explores how to help agents construct an Artificial Theory of Mind (AToM) by providing quantified, machine - readable prior information about human team members. This method aims to capture the characteristic profiles of individual team members, and these profiles can be constructed from sparse data and then provided to agents to support the development of their Artificial Theory of Mind. In addition, this research also explores how these profiles capture the team - composition characteristics that may affect team performance. The research examines through an experiment how these factors affect the performance of ad - hoc teams composed of humans and Artificial Social Intelligence (ASI) when performing complex team - coordination tasks. The experimental results show that the profiles of individuals and emerging teams can represent the team characteristics that predict behavior and explain the different perceptions of ASI. In particular, the analysis shows that for low - potential teams, ASI advisors have a significant positive impact and can improve the performance of these teams on task - outcome indicators. Overall, this paper aims to improve team effectiveness and social - cognitive levels in human - machine collaboration by developing the Theory - of - Mind capabilities of artificial intelligence.