What Intelligent Team Tutoring Systems Can Learn from Human-Agent Teams

Intelligent tutoring systems, having been relatively successful at emulating the results of human tutors for certain learning domains, are now being developed for team tutoring. With the addition of multiple humans in the system, modeling the communication and coordination between the humans and the tutoring agent grows combinatorically and present significant challenges to team tutor development. Some answers to these challenges can be found in the research field of human-agent teamwork, which is not typically followed by the intelligent tutoring systems research community. This paper applies common concepts in human-agent team literature, such as task allocation, adaptive automation triggers, and behavior modeling, to steps used to author a team tutor. This research should enable developers of intelligent team tutoring systems (ITTSs) to draw more efficiently on research from two otherwise siloed research areas.