Collaborative research: CER: reps-AI: supporting students in using genai in responsible, effective, and pedagogically sound ways for computational coursework across disciplines

The National Science Foundation Division of Computing and Communication Foundations awarded Carleton College $267,975 on October 1, 2026, to implement and evaluate tutoring-based models for teaching generative AI coding skills to non-computing majors across three institutions of higher education.

The project explores how higher education can support students in using generative AI tools in responsible, effective, and pedagogically sound ways for computational coursework. Carleton College, a small liberal arts institution, collaborates with two large public research universities to test three distinct approaches for providing campus-wide REPS-AI (Responsible, Effective, Pedagogically Sound AI) support. The three participating institutions—ranging in size and type—will implement different models adapted to their respective resources, structures, and strategies. By comparing the effectiveness of these tutoring approaches, the project will develop generalizable insights applicable to a broad range of non-computing students across U.S. higher education who need to develop essential generative AI coding skills but receive insufficient institutional support.

Work is performed at Carleton College in Northfield, Minnesota through September 30, 2029. The award is funded under the Computer and Information Science and Engineering program (CFDA 47.070). This is assistance funding classified as a Project Grant, supporting investigator-initiated research and education in computing and information science and engineering.

Abstract

Generative artificial intelligence (genai) is changing how computer code is being written. Consequently, IT is having dramatic impacts on education in computer science and other technological fields. But how will IT impact the rest of campus? Many scholars across campus use coding in their work, so coursework for computational scientists, computational artists, and digital humanities scholars (among others) has always involved coding. These students will need different kinds of genai support than future professional software developers, because their tasks and their fields are different. This project's goal is to explore a range of tutoring-based models for teaching AI skills to non-computing majors, implemented at three different institutions of higher education. By comparing the effectiveness of these models, the project will both develop generalizable insights into AI tutoring methods and have direct impacts across a wide range of students in the U.S. The project is grounded in the observation that higher education is not equipped to support all students on campus in developing essential genai coding skills. The central goal of the project is to lay the intellectual and practical foundations to assist institutions in supporting all students in using genai tools in responsible, effective, and pedagogically sound (reps-AI) ways to accomplish their computational coursework. Because different institutions will provide reps-AI support in different ways based on their resources, structures, and strategies, this proposal will implement and evaluate three distinct approaches for providing campus-wide reps-AI support to students at the three different schools. The three participating institutions range in size (two large, public, R1 land-grant institutions; one small liberal arts college) and hold differing visions for the role of genai on campus, ranging from early adoption/widespread integration to cautious and strategic approaches. Each campus also has differing existing infrastructure for providing reps-AI support, resulting in three distinct strategies being pursued, including developing a new campus-wide student support center, integrating reps-AI supports to existing student support centers, and creating mini-courses and workshops for students across campus. This project will produce three models of reps-AI student supports, case studies of how to establish each on an IHE campus, and comparative analysis of the approaches. This in turn will allow other institutions to more quickly and confidently use the insights gained here to improve their own AI educational efforts. 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.