• Jul 10, 2018

Back to the Future of Pedagogical Agents

Pedagogical agents are becoming personalized, lifelong learning companions

The recent paper “Pedagogical Agents: Back to the Future” by Dr. Lewis Johnson of Alelo Inc. and Professor James Lester of North Carolina State University reviews the history and likely future of pedagogical agents, which are autonomous, interactive computer-based characters that engage in rich, face-to-face interactions with learners.

Screenshot of VCAT Taiwan.

The authors propose that fast-changing workplaces in the global economy require lifelong learning of new skills, which drives the need for more cost-effective learning solutions. Instructor-based learning is cost-prohibitive for most applications, so learning solutions are moving toward digital ecosystems populated by ‘smart’ pedagogical agents based on advances in AI, natural language processing, machine learning, and virtual and augmented reality. A key outcome of this trend will be that pedagogical agents will become learners’ personalized, lifelong companions.

In the big scheme of things, AI-based learning methods are fundamentally transforming the global training and education industry by transcending the traditional boundaries of educational institutions.

The recent paper follows the authors’ groundbreaking paper “Animated Pedagogical Agents: Face-to-Face Interaction in Interactive Learning Environments” published by the International Journal of Artificial Intelligence in Learning in 2000. The authors proposed then that pedagogical agents had the potential to engage learners and promote better learning with interactive, humanlike ways such as verbal and nonverbal feedback, gesture, and emotional expressions. The paper’s contributions to learning and education merited its receipt of the prestigious IFAAMAS Award for Influential Papers in Autonomous Agents and Multiagent Systems in 2017.

This year’s paper presents validation of the prediction made 18 years ago that pedagogical agents can measurably improve learning relative to modalities that do not use agents. A key finding is that the effects of pedagogical agents vary greatly depending on subject matter and learner population. For example, role-playing pedagogical agents are highly valuable for learning languages, cultures, and soft skills. An early example of this application is Alelo’s Tactical Language & Culture Training that has been used by tens of thousands American and Australian military members since 2003.

Alelo’s Dr. Johnson will present insights on the value and technology of pedagogical agents during his webinar “How Will AI and Data-Driven Learning Transform the Global Education Industry?” sponsored by the SIIA.

—————————————————————————–

In 2000 the International Journal of Artificial Intelligence in Learning published a groundbreaking paper on pedagogical agents, then a very novel paradigm in interactive learning. Pedagogical agents are autonomous computer-based characters that engage in rich, face-to-face interactions with learners.

The paper’s authors — Lewis Johnson and Jeff Rickel of the University of Southern California, and James Lester of North Carolina State University — proposed that pedagogical agents had the potential to engage learners and promote better learning with interactive, humanlike ways such as verbal and nonverbal feedback, gesture, and emotional expressions.

Eighteen years later, research has validated these predictions in that pedagogical agents measurably improve learning relative to modalities that do not use agents. An early example of this application is Alelo Inc.’s Tactical Language & Culture courses used by thousands of members of the U.S. and Australian militaries since 2003. The paper’s contributions to learning and education technologies merited its receipt of the prestigious IFAAMAS Award for Influential Papers in Autonomous Agents and Multiagent Systems in 2017.

Dr. Johnson, now CEO of Alelo Inc., and Professor Lester recently published the continuation article “Pedagogical Agents: Back to the Future” to review what has happened in pedagogical agents since the year 2000, and make new predictions. A key finding is that the effects of pedagogical agents vary greatly depending on subject matter and learner population. For example, pedagogical agents are invaluable for teaching languages, cultures, and soft skills by role-playing the target skills with learners.

The authors propose that the global economy increases the demand for lifelong learning of new skills, which drives the need for more cost-effective learning solutions. As instructor-based learning continues to be cost-prohibitive for most applications, learning solutions move toward online learning and digital ecosystems populated with ‘smart’, AI-based pedagogical agents made possible by advances in natural language processing, machine learning, and virtual and augmented reality.

As learning transcends the traditional boundaries of educational institutions, AI-based learning methods are fundamentally transforming the global training and education industry. An important outcome will be that pedagogical agents will be learners’ personal lifelong companions.