Paper Review: Research-Based Design of Pedagogical Agent Roles: a Review, Progress, and Recommendations Yanghee Kim & Amy L. Baylor

Paper link here

Original Keywords : Pedagogical agents . Experimental research .Virtual humans.Interface design
My Keywords : Motivating learning , pedagogical digital agent, human-agent relationship


The paper caught my attention because it was in the context of educational tools and promised to have practical recommendations on designing pedagogical agents.

In summary they have designed a learning environment and in an experimental setting expose learners to 3 different artificial pedagogical agents: mentor, expert and motivator. Then they measured the students' motivation and learning outcomes plus their perception of agent's character(did they perceive them as intended by designer).


It related to the Reeves and Nass's work they found several social and human-like relation between students and agnates. For example students preferred agents with the same race as themselves, they felt motivated or hurt with agents comments.
They also discovered that agent's intelligence is not a must, for having a human-like experience with agent and improved learning outcomes: "The pre-scripted agents in our original study, however, effectively played their intended instructional roles without being particularly intelligent."(Kim & Baylor, 2016)
In addition, they found out that social and human characteristics were assigned to agent by students. for example expert was seen as more credible.


practical suggestions:
Split-persona effect: In this work they found out that two agents each with its distinct persona is more effective than one agent that has both functionalities. For example having "expert" and "motivator" works better than having only "mentor"agent that is "supposedly"(because I dont know how they actually implemented this persona) combines those two qualities. The rationale as they put it : "It seemed that it was easier for students to figuratively compartmentalize the agent information when it was delivered by two distinct sources."(Kim & Baylor, 2016)

Kim, Y., & Baylor, A. L. (2016). Research-based design of pedagogical agent roles: a review, progress, and recommendations. International Journal of Artificial Intelligence in Education, 26(1), 160-169.