Smart (but Also) Challenging Learning Environments: The Case of Conversational Agents That Foster Productive Peer Dialogue

Abstract

Smart learning environments (SLEs) are typically described as environments that offer rich learning interactions by making use of digital, context-aware, and adaptive technologies. This work introduces the concept of “smart challenging learning environments” (SCLEs), referring to SLEs that employ computational models of a “learner challenging” strategy. Such a strategy is conceptualized as a didactical model that enables the system to initiate learner-oriented interventions aiming to generate new challenges as opportunities for further learning. System interventions are expected to heighten students’ interest and direct their focus, thus resulting in productive cognitive activity and improved learning outcomes. To justify this approach, the learner “challenging” vs. “accommodating” approach is discussed contrasting research evidence from some widely adopted learning design paradigms, such as the “learning styles” and “scripted collaboration.” Next, the focus turns on the conversational agent domain analyzing the design and learning impact of “MentorChat,” an agent-based system that implements a learner challenging strategy, based on the discourse framework of academically productive talk (APT), to support and trigger student dialogue in web-based collaborative learning settings. Evidence from three studies is provided presenting the learning benefits and shortcomings that emerge when the MentorChat agent challenges peers to deeper discuss aspects of the learning domain. Finally, MentorChat-based conclusions are generalized toward providing system requirements for developing Smart and Challenging Learning Environments. The implementation of a learner challenging strategy is proposed as a key requirement in SCLEs while labor-efficient development is suggested through keeping system modeling at a minimum level. Teacher-friendliness is also emphasized through enabling system configuration by the teacher end-user.