Collaborative research: augmenting self-regulated learning through human-AI CO-regulation

The National Science Foundation's Office of Multidisciplinary Activities awarded the Regents of the University of Michigan $237,500 on September 1, 2026, through the Social, Behavioral, and Economic Sciences program (CFDA 47.075) to develop and evaluate an AI-powered learning support system that augments student self-regulated learning through human-AI co-regulation.

The project develops artificial intelligence models that track student mastery of course concepts and generate personalized practice questions, then designs system interfaces that make the AI's reasoning transparent and controllable to students. Students can inspect, reflect on, challenge, and revise the system's estimate of their concept mastery, aligning their self-assessment with the system's computational estimate. The research team works iteratively with instructors and students to design the system interfaces, then implements and evaluates the system in authentic courses. Evaluation includes analysis of AI-generated practice questions, system interaction records, and student outcomes.

Work is performed at the University of Michigan in Ann Arbor, Michigan, with a period of performance from September 1, 2026, through August 31, 2029. The assistance type is a project grant.

Abstract

Artificial intelligence (AI) is rapidly changing how students learn by offering tools that can estimate what a student knows and suggest what to study next. These tools can be helpful, but they often make decisions in ways that students cannot understand, which may leave students feeling less in control of their own learning. This project studies how structured interaction between learners and artificial intelligence (AI) systems can augment human cognitive regulation in authentic learning environments. Specifically, the project develops and studies a novel AI-powered learning support system that shows students its understanding of what they have and have not mastered, and allows them to inspect, reflect on, challenge, and revise that understanding. By making technology transparent and controllable, the project helps students remain active partners in their learning rather than passive recipients of automated support. The research team aims to work with instructors and students to iteratively design this system together and evaluate its impact on student-AI collaboration in supporting their learning. This project aims to investigate how students and AI can CO-regulate learning through a system that transparently presents its estimate of a student?S concept mastery as a visible representation that the student can reflect on and revise, aligning the student?S self-assessment with the system?S computational estimate. In aim 1, the research team develops AI models that track students? Mastery of course concepts and generate personalized practice questions based on that evolving estimate. In aim 2, researchers work with instructors and students to design system interfaces for this new AI system. Finally, aim 3 is focused on implementation and evaluation of the system in authentic courses. The team analyzes AI-generated practice questions, system interaction records, and students? Performance on final exams towards producing generalizable knowledge about how cognitive control is distributed across humans and AI during learning. This project builds foundational knowledge about how transparent and controllable AI can support student agency, cognitive regulation, and learning. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.- subawards are not planned for this award.