Conversational Agents as Group-Teacher Interaction Mediators in MOOCs

Abstract

This paper presents the colMOOC project perspective for integrating a conversational agent service in Massive Open Online Courses (MOOCs) to facilitate peer dialogue during chat-based activities. Furthermore, it provides high level system architecture to highlight key aspects of the technology infrastructure and the functionality of the software system. The rationale of the approach follows the tenets of several converging theoretical perspectives emphasizing the value of learners' dialogue and constructive argumentation as knowledge building processes during collaborative activities. The MOOC teacher (domain expert) is expected to play a key role by entering in the system conceptual links that guide the agent to make relevant questioning interventions. Thus, the agent tool is mediating group-teacher interactions in a MOOC environment aiming to increase the transactive quality of peer dialogue. The paper discusses also the `Academically Productive Talk' model that provides the basis for the agent interventions and the major software components of the system, including the colMOOC Editor and the colMOOC Player.