AI as the Coding Partner: Instructional Practices for Artificial Intelligence Pair Programming in High School Computer Science Courses

About This Grant

Advances in artificial intelligence (AI) are having profound impacts in a range of areas, therefore computer science (CS) education must prepare students to effectively use AI tools. Because these tools are so new, best practices for educational use of AI tools are not well understood; in fact, AI tools may actually inhibit student learning if not used well. Also, because a majority of professional software developers now use AI tools, it is important for students to develop proficiency in using these tools - but it is also important that students use them in a way that enhances their learning instead of replacing it. This project contributes to the effective integration of AI tools in CS education by investigating pair programming with AI (AIPP). In traditional pair programming, two students work collaboratively to complete an assignment. AIPP involves one student engaging in pair programming with an AI tool. The project is developing a set of curricular materials to support students to productively use AI in their learning to code. The materials are curriculum-agnostic--that is, they are an "overlay" to be used along with any other high school CS curriculum. The materials direct students on how to complete curricular tasks so that their prompts to AI and what they do with AI responses are supportive of their learning. In this way, the project is supporting students to use AI efficiently as a learning tool, while at the same time mimicking ways that the work of industry in CS is adopting AI. As such, the project is exploring ways to prepare students for introduction into the workforce while leveraging AI as an efficient tool to bolster CS education generally. This project is researching the knowledge and skills students need in order to successfully engage in AIPP. Building on early research of promising practices for educational AIPP, the project studies the efficacy of two recommendations: (1) developing skills for code reading and review, and (2) modulating trust in AI. The project is particularly focused on uncovering evidence of best practices that support all students in their CS learning, regardless of their prior CS experience. The first phase of the project is developing AIPP resources with CS educators in Region 8 of the STEM Collective for Innovative Louisiana Stakeholders. Following this, CS educators implement AIPP in their classrooms. Repeated measures designs test the efficacy among students of AIPP (and the two supportive practices) relative to traditional pair programming. The project is iteratively revising and updating the curricular materials after the first round of implementation and then preparing additional teachers to implement AIPP for further testing in two more years. Both the research findings and the curricular materials will be disseminated to various appropriate audiences, such as other researchers, teachers and curriculum designers. This project is funded by the CS for All: Research and RPPs program. 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.