Proposing an AI-Assistant to Help Teachers Analyse and Modify Computational Problem Solving Activities
Year: 2026
Type: conference-paper
Abstract: With an increasing amount of resources for teaching Computational Problem Solving (CPS) in K–12, teachers must be able to critically analyse programming activities to determine which to select and how to modify them to foster CPS effectively. High-quality CPS activities present problems that allow meaningfully different solution strategies and offer appropriate cognitive challenge, yet these features are difficult to evaluate in practice. To explore whether AI tools could support this process, we developed a CPS activity analysis guide that operationalises these core characteristics and used it to configure the prototype of an AI-assistant through detailed prompt specification. We applied the assistant to eight freely accessible programming activities from four platforms and compared its analyses with those of CS education experts. The AI-assistant largely aligned with, and often complemented, expert judgments in identifying meaningfully different solutions and in suggesting ways to adjust CPS demand. However, the joint analyses also revealed structural limitations of many online activities (namely particularly strict task constraints and non-configurable environments) which restrict teachers’ ability to implement modifications that would increase CPS demand. These findings highlight the promise of AI-supported CPS activity analysis, and future work will continue to refine the guide and AI-assistant, and test with a wider range of activities and practitioners, including teachers.
Institution: Haute École Pédagogique du Canton de Vaud
Language: English
Subfield: Computer Science Applications
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: hybrid
OpenAlex ID: W7203837109