Pedagogical Agents: Back to the Future
W. Lewis Johnson,James C. Lester
DOI: https://doi.org/10.1609/aimag.v39i2.2793
IF: 0.9
2018-07-01
AI Magazine
Abstract:Back in the 1990s we started work on pedagogical agents, a new user interface paradigm for interactive learning environments. Pedagogical agents are autonomous characters that inhabit learning environments and can engage with learners in rich, face-to-face interactions. Building on this work, in 2000 we, together with our colleague, Jeff Rickel, published an article on pedagogical agents that surveyed this new paradigm and discussed its potential. We made the case that pedagogical agents that interact with learners in natural, life-like ways can help learning environments achieve improved learning outcomes. This article has been widely cited, and was a winner of the 2017 IFAAMAS Award for Influential Papers in Autonomous Agents and Multiagent Systems (IFAAMAS, 2017). On the occasion of receiving the IFAAMAS award, and after twenty years of work on pedagogical agents, we decided to take another look at the future of the field. We’ll start by revisiting our predictions for pedagogical agents back in 2000, and examine which of those predictions panned out. Then, informed what we have learned since then, we will take another look at emerging trends and the future of pedagogical agents. Advances in natural language dialogue, affective computing, machine learning, virtual environments, and robotics are making possible even more lifelike and effective pedagogical agents, with potentially profound effects on the way people learn.
computer science, artificial intelligence