TY - CHAP
T1 - Conversational agents in MOOCs
T2 - reflections on first outcomes of the colMOOC project
AU - Demetriadis, Stavros
AU - Caballe, Santi
AU - Papadopoulos, Pantelis M.
AU - Gómez-Sánchez, Eduardo
AU - Kolling, Allison
AU - Tegos, Stergios
AU - Tsiatsos, Thrasyvoulos
AU - Psathas, Georgios
AU - Michos, Konstantinos
AU - Weinberger, Armin
AU - Winther Bech, Christian
AU - Karakostas, Anastasios
AU - Tsibanis, Costas
AU - Palaigeorgiou, George
AU - Hodges, Matthew
PY - 2021/6/8
Y1 - 2021/6/8
N2 - Massive Open Online Courses (MOOCs) provide a powerful means for informal online learning that is already popular, engaging great numbers of students all over the world. However, studies on MOOCs’ efficiency frequently report on the high dropout rates of enrolled students, and the lack of productive social interaction to promote the quality of MOOC-based learning. Conversational agents (CAs), on the other hand, appear to be a promising artificial intelligence (AI) technology with the potential of acting as catalysts of students’ social interaction, a factor known to beneficially affect learning at many levels. Within this context the “colMOOC” project has proposed and developed an agent-based tool and methodology for integrating flexible and teacher-configurable CAs along with relevant learning analytics services in MOOCs platforms, aiming to promote peer learning interactions. This chapter concisely presents the overall rationale of the project, the agent-based tools developed, and provides reflections on the first project outcomes emerging from four different pilot MOOCs offered so far. Early conclusions analyze the challenges for integrating a teacher-configured agent-chat service in MOOCs, provide helpful guidelines for efficient task design, and highlight promising evidence on the learning impact of participating in agent-chat activities.
AB - Massive Open Online Courses (MOOCs) provide a powerful means for informal online learning that is already popular, engaging great numbers of students all over the world. However, studies on MOOCs’ efficiency frequently report on the high dropout rates of enrolled students, and the lack of productive social interaction to promote the quality of MOOC-based learning. Conversational agents (CAs), on the other hand, appear to be a promising artificial intelligence (AI) technology with the potential of acting as catalysts of students’ social interaction, a factor known to beneficially affect learning at many levels. Within this context the “colMOOC” project has proposed and developed an agent-based tool and methodology for integrating flexible and teacher-configurable CAs along with relevant learning analytics services in MOOCs platforms, aiming to promote peer learning interactions. This chapter concisely presents the overall rationale of the project, the agent-based tools developed, and provides reflections on the first project outcomes emerging from four different pilot MOOCs offered so far. Early conclusions analyze the challenges for integrating a teacher-configured agent-chat service in MOOCs, provide helpful guidelines for efficient task design, and highlight promising evidence on the learning impact of participating in agent-chat activities.
U2 - 10.1016/B978-0-12-823410-5.00001-2
DO - 10.1016/B978-0-12-823410-5.00001-2
M3 - Chapter
SN - 978-0-12-823410-5
SP - xxxvii-lxxiv
BT - Intelligent Systems and Learning Data Analytics in Online Education
PB - Elsevier
ER -