Project-based learning is increasingly becoming an essential pillar in computing education (Pucher & Lehner, 2011). However, one major challenge of project-based learning is that students are expected not only to develop their computational thinking abilities but also to execute the full lifecycle of a project, emerging as effective project creators and managers (Pucher & Lehner, 2011). Thus, a range of tacit skills related to project planning, scoping, and execution come into play when integrating project-based learning into computational thinking pathways (Shin et al., 2021). One of the key skills and outcomes that educators aim to nurture and instill through the project-based learning approach is students’ ability to clearly articulate their goals, strategically plan their approaches, and adapt their objectives as they work toward executing a comprehensive project vision (Rachman et al., 2020). Furthermore, given the increasing use of team-based, collaborative modalities in today’s STEM workforce, training students to be well prepared to work in such environments is crucial (Rachman et al., 2020).
Beyond the ability to think and connect computational thinking to the different layers of planning and executing a project, project-based learning involves developing a structured motivational regime. This helps students stay engaged with the project while also balancing their varied academic, social, and personal goals (Kim et al., 2021). For example, when planning a project for a programming course, a student needs to factor in the time they will need to allocate to other courses and social obligations, such as student groups, and time that might be required to learn and adjust goals as needed during project development. Developing motivation and sustaining a stable rhythm for sustained progress are vital to the success of project-based learning. Yet developing sustained motivation can be challenging; thus, finding ways to help students adopt a sustained approach to achieving their project execution goals is important. While previous studies have explored the relationship between computational thinking and the goals of project-based learning (Rachman et al., 2020), the mechanism by which students regulate their motivation throughout the process remains unclear. Our proposed work contributes to this vision by developing and testing a theoretically grounded tool that utilizes generative Artificial Intelligence (AI) to support both motivational regulation and project planning.
Various generative AI tools (such as ChatGPT) are increasingly part of the ecological landscape of instructional planning and execution (Lo, 2023). These are reshaping the way students learn and are emerging as an essential part of the toolkit educators possess to innovate classroom experiences and enhance the novelty of instruction (Lo et al., 2021). This further motivates the premise of designing a tool that not only provides instructional support but also facilitates planning and execution of projects, drawing on the framework of multiple goals regulation (Kim et al., 2023). The key questions this study will explore are:
1) Can an AI assistant provide necessary motivational cues to help manage and encourage students for project-based learning in computing education?
2) How might educators perceive the impact of such tools on learning outcomes and classroom dynamics?
3) How might such a tool impact the self-efficacy and creative ideation of students?
The first stage of this project will involve developing a tool using generative AI that is also aligned with the Multiple Goals Regulation framework (Kim et al., 2023). The tool is currently in development and testing. The goal is to develop a context-aware system that leverages the generative capabilities of GPT models (akin to ChatGPT) while customizing its responses based on theoretical and empirical foundations from prior literature (e.g., Kim et al., 2021, 2023). To achieve this, we use the Retrieval Augmented Generation pipeline implemented in Python (Ke et al., 2024). Initial tests with this tool highlight promising directions, demonstrating an AI’s ability to provide responses aligned with motivational processes, such as the effective delineation of sub-goals and strategies for long-term goal planning. These capabilities can help students and educators better structure and manage project-based learning, particularly in computing education. Using the naive ChatGPT model as a baseline, we found that a customized model provides a greater understanding of the focus on motivation and planning of a project, whereas a generic GPT model primarily provides a broad overview of computational aspects (such as selection of datasets, standard algorithms) without acknowledging the implicit motivational challenges that may impair the effectiveness of project-based learning. This further highlights the need to develop a tool grounded in conceptual and empirical evidence, as this ensures the design of an AI system that can offer students meaningful support in regulating their motivation and ultimately succeed in their project-based learning context and beyond.
Future work will entail refining and scaling the initial prototype. In particular, various design workshops and survey studies will help enrich the ways stakeholders would like to interact with and collaborate with such an AI system. These efforts will be pivotal in guiding the tool’s development and ensuring it effectively meets user needs and aligns with prior literature. Subsequent analyses will focus on incorporating the voices and expectations of educators and students regarding how this tool should interact and the level of agency it should possess, so that it enriches rather than diminishes the effectiveness of instruction. Once the tool is fully functional, our goal will be to conduct a longitudinal, field-based experiment in a computer science class to understand how students use this tool, how it impacts and drives evolution in their learning strategies and goals, and to ultimately collect feedback from educators. Our work will provide valuable insights into the role of AI-based tools in computational coursework, highlighting their potential in supporting both computational thinking and soft skills (pivotal for developing a sustained motivational pattern), as well as the strategies needed to thrive in the workforce.
EdTech Archives