AI-driven code critique across the undergraduate computing curriculum

The National Science Foundation Division of Undergraduate Education awarded Michigan Technological University $599,238 on October 1, 2026, under the STEM Education program (CFDA 47.076) to develop an adaptive, scalable AI-enhanced code critique system for identifying coding antipatterns across the undergraduate computing curriculum. The project, designated as a Level 2 Engaged Student Learning initiative, runs through September 30, 2029, with performance at Houghton, Michigan.

The system will deliver contextualized, timely feedback on student code to promote genuine coding proficiency and counter student reliance on AI-generated solutions. The project addresses three research questions: the effectiveness of AI-enhanced code critiquing in identifying antipatterns and defects; how the system's performance and behavior adjust to different educational contexts; and the impact of AI-driven code critiques on student learning outcomes and behaviors. Effectiveness will be assessed using precision and recall metrics on antipattern detection, while learning impact will employ pre- and post-surveys, analysis of student code and system interactions, and qualitative feedback. Upon completion, the developed infrastructure and comprehensive deployment guides will be made available to facilitate adoption across diverse educational contexts.

Abstract

This project aims to serve the national interest by developing an adaptive, scalable AI-enhanced code critiquer system capable of identifying coding antipatterns across the computing curriculum. Antipatterns are recurring poor solutions that novice coders make. This project will strengthen the nation?s workforce by incorporating a code critiquer system in undergraduate computing curriculum to provide appropriate, contextualized, and timely feedback on student?s code in order to promote genuine coding proficiency rather than reliance on AI-generated solutions. This level 2 engaged student learning project?s insights and outputs will enhance current educational practices and prepare students for the evolving demands of the computing workforce. Upon project completion, the developed infrastructure and comprehensive deployment guides will be available and facilitate adoption across diverse educational contexts. The project will explore three key research questions: (1) how effectively can an AI-enhanced code critiquing system identify code antipatterns and defects across the undergraduate computing curriculum? (2) how does the AI-driven critiquer?s performance and behavior adjust to different contexts? (3) what impact does AI-driven code critiques have on student learning outcomes and behaviors? Each research question will be assessed using different metrics and proven methodologies. For example, the effectiveness of the AI system will use precision and recall on the ability to detect antipatterns. To evaluate impact on student learning, the project will employ pre- and post- surveys, analysis of student code and interactions with the system, and qualitative feedback, etc. The project addresses current pedagogical disruptions involving student reliance on generative AI, promoting authentic learning experiences and improving student engagement and confidence. The NSF iuse: EDU program supports research and development projects to improve the effectiveness of stem education for all students. Through the engaged student learning track, the program supports the creation, exploration, and implementation of promising practices and tools. 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.