*Liu X. & Pourdavood, R. G. (2026). Designing for AI-driven self-directed learning in math: A mixed methods study of pre-service teachers’ emerging competence. Paper presented at the 53rd Annual Meeting of the Research Council on Mathematics Learning (RCML), Las Vegas, Nevada, March 5-7.

This study explores the impact of an AI-driven self-directed learning intervention on pre-service teachers’ attitudes toward and competence in using AI for mathematics education. Participants, enrolled in two math methods courses at a Midwestern university, will engage in iterative learning cycles involving Blackboard-based assessments and generative AI support via ChatGPT. After receiving immediate feedback on missed questions, participants will use ChatGPT to access explanations, worked examples, conceptual representations, and real-world applications before retaking equivalent tests to achieve mastery. Data collection will include pre- and post-surveys, interviews, and reflective essays to capture both cognitive and affective dimensions of learning. Grounded in Garrison’s (1997) framework of self-directed learning—self-management, self-monitoring, and self-motivation—the study examines how AI can enhance learner autonomy and metacognitive engagement. Findings are expected to inform teacher education programs, address critical gaps in AI literacy, and contribute to ongoing debates about integrating generative AI into K–12 mathematics instruction.