With the increasing popularity of AI and machine learning, human-AI teaming has a wide range of applications in transportation, healthcare, the military, manufacturing, and people’s everyday life. Measurement of human-AI team effectiveness is essential for guiding the design of AI and evaluating human-AI teams. To develop suitable measures of human-AI teamwork effectiveness, we created a search and rescue task environment in Minecraft, in which Artificial Social Intelligence (ASI) agents inferred human teams’ mental states, predicted their actions, and intervened to improve their teamwork (Huang et al., 2022). As a comparison, we also collected data from teams with a human advisor and with no advisor. We investigated the effects of human advisor interventions on team performance. In this study, we examined intervention data and compliance in a human-AI teaming experiment to gain insights into the efficacy of advisor interventions. The analysis categorized the types of interventions provided by a human advisor and the corresponding compliance. The finding of this paper is a preliminary step towards a comprehensive study on ASI agents, in which results from the human advisor study can provide valuable comparisons and insights. Future research will focus on analyzing ASI agents’ interventions to determine their effectiveness, identify the best measurements for human-AI teamwork effectiveness, and facilitate the development of ASI agents.

Human team members show a remarkable ability to infer the state of their partners and anticipate their needs and actions. Prior research demonstrates that an artificial system can make some predictions accurately concerning artificial agents. This study investigated whether an artificial system could generate a robust Theory of Mind of human teammates. An urban search and rescue (USAR) task environment was developed to elicit human teamwork and evaluate inference and prediction about team members by software agents and humans. The task varied team members’ roles and skills, types of task synchronization and interdependence, task risk and reward, completeness of mission planning, and information asymmetry. The task was implemented in MinecraftTM and applied in a study of 64 teams, each with three remotely distributed members. An evaluation of six Artificial Social Intelligences (ASI) and several human observers addressed the accuracy with which each predicted team performance, inferred experimentally manipulated knowledge of team members, and predicted member actions. All agents performed above chance; humans slightly outperformed ASI agents on some tasks and significantly outperformed ASI agents on others; no one ASI agent reliably outperformed the others; and the accuracy of ASI agents and human observers improved rapidly though modestly during the brief trials.