Pedagogical agents with conversational dialogue are becoming more popular in today's advanced learning environments. The intelligence of the knowledge tracking, conversational moves, and other features of these agents has been improving over the last decade and skepticism about this technology is waning. This presentation covers the 12-year evolution of AutoTutor, a learning environment with an agent that tutors students by holding a conversation in natural language. AutoTutor presents challenging questions or problems and then engages in dialogue that guides the student in building an answer. There are versions of AutoTutor that guide interactive simulation in 3D microworlds, that detect and produce emotions, and that are embedded in games. AutoTutor improves learning over reading textbooks for an equivalent amount of time. The presentation covers AutoTutor and other learning environments with conversational pedagogical agents.
展开▼