2026 ASEE Annual Conference & Exposition

This Complete Evidence-based Practice paper examines how first-year engineering students engage with generative artificial intelligence (AI) tools to support Arduino coding in team-based design projects. As AI tools such as Copilot become increasingly integrated into programming environments, educators are exploring how to leverage them to support novice learners. Arduino-based projects provide a tangible platform for rapid prototyping and team collaboration. Yet as AI tools generate code, explain syntax, and troubleshoot hardware integration, questions arise about how interactions between students and AI tools reshape students’ learning, teamwork, and development of programming confidence.

Recent studies have begun to examine these dynamics. Barke et al. [1] identified two main interaction modes: acceleration, where programmers used Copilot to speed up tasks, and exploration, where they used it to discover options when unsure how to proceed. Prather et al. [2] further described two interaction patterns: shepherding, where students guided Copilot to generate useful code, and drifting, where students followed incorrect suggestions and became lost, highlighting the need for explicit instructional scaffolding. While studies reported significant increases in students’ familiarity and use of AI tools, concerns about academic integrity and over-reliance persisted [3]. Research exploring learning outcomes found that novice students who wrote their own Arduino programs developed greater self-efficacy and programming ability compared to those using ChatGPT [4].

Despite growing interest, several gaps remain. Most research on AI-assisted programming focuses on individual learners in text-based environments rather than on teams engaged in open-ended, physical computing projects. Few empirical studies have examined structured scaffolds such as informed appropriate use, reflection, and ethical discussion that guide students to engage with AI tools effectively and responsibly in project-based learning. Moreover, how AI tools affect team dynamics such as communication, role shifts, and collective reasoning remains largely unexamined. Addressing these gaps is critical for informing evidence-based integration of generative AI into first-year engineering curricula.

At our institution, first-year engineering students in Engineering Foundations I (fall semester) undertake team projects that combine Arduino programming, circuitry, and 3D design in the first half of the term. Because students’ experience levels in both coding and Arduino vary widely, integrating generative AI presents a promising opportunity to supplement and support their learning of Arduino coding. This study addresses three questions: (1) How do students leverage generative AI for Arduino coding in their team-based projects? (2) How do students consider ethical implications of AI use? (3) How does AI use affect team dynamics?

The research uses a pre-post survey and reflection approach, collecting data from first-year students in Engineering Foundations I. Baseline data captured students’ pre-college experience with AI tools, including purposes of use, reliability of information, confidence, and perceived advantages and limitations. The post-study survey prompted students to reflect on their AI use and its impact on learning including the extent of AI involvement in code generation, the specific ways AI was used, students’ evaluations of how AI supported their learning, ethical considerations, and effects on teamwork.

Over the past two years, AI-assisted Arduino coding was introduced with iterative refinement. In the first year, instruction emphasized prompting strategies, proper citation of AI tools, and demonstrations using simple flashing LED examples. Students struggled with ethical and practical considerations, reviewing the need for more explicit scaffolding. In the second year, we developed and implemented an AI-Assisted Arduino Coding Guide that covered effective prompting, when and how to use AI in Arduino coding without compromising learning, how to cite AI tools, and ethical considerations across seven dimensions, along with best practices and tips. We also incorporated case studies on AI ethics to promote discussion and reflection. These materials and efforts aimed to contextualize responsible use of AI tools in coding practice in project-based learning.

Preliminary analyses of the baseline data indicated that nearly all students had prior experience with AI tools before college, with most using them weekly or monthly. Fewer than half had used AI for coding or debugging. Students reported moderately to high confidence in using AI, citing its usefulness for supporting their learning, while also expressed concerns about inaccuracy and overreliance. Post-study analysis, to be completed later in the semester, will examine changes in students’ confidence, ethical awareness, learning experiences with AI-assisted Arduino coding, and perceptions of how AI influenced teamwork.

This research contributes evidence to designing structured interventions that cultivate responsible AI use, ethical reasoning, and collaborative problem-solving in engineering design education. Findings will inform strategies for integrating AI tools as cognitive and social scaffolds that enhance, rather than replace, students’ foundational programming skills and creative problem-solving in team-based contexts.

References
[1] S. Barke, M. B. James, and N. Polikarpova, “Grounded Copilot: How Programmers Interact with Code-Generating Models,” Proc. ACM Program. Lang., vol. 7, no. OOPSLA1, pp. 85–111, Apr. 2023, doi: 10.1145/3586030.
[2] J. Prather et al., “‘It’s Weird That it Knows What I Want’: Usability and Interactions with Copilot for Novice Programmers,” ACM Trans. Comput.-Hum. Interact., vol. 31, no. 1, pp. 1–31, Feb. 2024, doi: 10.1145/3617367.
[3] R. Zviel-Girshin, “The Good and Bad of AI Tools in Novice Programming Education,” Educ. Sci., vol. 14, no. 10, p. 1089, Oct. 2024, doi: 10.3390/educsci14101089.
[4] D. Johnson, W. Doss, and C. Estepp, “Using ChatGPT with Novice Arduino Programmers: Effects on Performance, Interest, Self-Efficacy, and Programming Ability,” J. Res. Tech. Careers, vol. 8, no. 1, p. 1, May 2024, doi: 10.9741/2578-2118.1152.