Adaptive Motivational Support in LLM-Based Pedagogical Agents: Design and Evaluation.

Large language models (LLMs) create opportunities for pedagogical agents to move beyond answering questions and provide responsive motivational support during learning. However, generating fluent or encouraging dialogue does not establish that an agent helps learners sustain effort, recover from setbacks, or make autonomous learning decisions. A central design challenge is to determine when support is needed and how its content, timing, and intensity should adapt to learner goals, interaction histories, and task contexts. This Research Topic focuses on adaptive motivational support delivered through learner-facing, LLM-based pedagogical agents. Rather than surveying AI in education broadly, it examines the design and evaluation of specific interaction mechanisms that connect learner needs, motivational strategies, and educationally meaningful outcomes.The goal is to advance understanding of how LLM-based pedagogical agents can deliver motivational support that is both responsive and pedagogically grounded. We seek contributions that specify the motivational problem being addressed, the information used to adapt support, and the role of the LLM in generating or selecting responses. Relevant mechanisms include tailored encouragement after setbacks, feedback supporting learners’ sense of competence, dialogue about task value and goals, and adjustments to the frequency or directiveness of assistance. Contributions should examine how these mechanisms affect learner motivation, with engagement, persistence, self-regulation, and learning outcomes considered where relevant. Particular attention is given to supporting learner choice and avoiding excessive prompting or or dependence on the agent. Both controlled studies and classroom or informal-learning deployments are welcome. The collection aims to produce actionable design principles and evaluation methods, without assuming that more interaction or longer system use necessarily represents better learning.We welcome Original Research, Methods, Technology and Code articles, and focused Systematic Reviews addressing adaptive motivational support in LLM-based pedagogical agents. Themes include:• Pedagogically grounded dialogue strategies for motivational scaffolding.• Learner modeling and context-sensitive adaptation of support content, timing, or intensity.• Comparisons of adaptive and non-adaptive support, including effects on learner autonomy.• Evaluation of motivational and learning-related outcomes across single or repeated interactions.Contributions must address a specific motivational-support mechanism in a learner-facing LLM-based agent, or synthesize evidence directly relevant to such mechanisms. Hybrid and multi-agent architectures are within scope only where they advance this focus. General AI adoption surveys, generic educational content generation, broad AI ethics discussions, and health-promotion or lifestyle-change interventions are outside scope. Topic editors: Mic Jiang, Wei Miao Source: https://www.frontiersin.org/research-topics/83476/adaptive-motivational-support-in-llm-based-pedagogical-agents-design-and-evaluation