A Constructionist Approach to Learning with AI: An “Old” Solution to a “New” Problem?

Abstract

The advent of large language models has sparked significant discussions within the academic community. It has enabled students to effortlessly generate near-flawless essays, prompting varied responses from academics ranging from calls for a complete ban to a fundamental re-evaluation of teaching and assessment practices. This paper presents a case study that demonstrates the creation of a three-part learning task, enabling students to utilize Artificial Intelligence (AI) tools to produce high-quality work while employing an evaluation rubric capable of assessing the output regardless of AI usage. The design experiment used various teaching tools and methods to facilitate the systematic literature review process. The resulting student artifacts were visually analyzed and compared against the assessment rubrics to identify strengths and weaknesses. Students openly employed AI in generating their literature reviews, emphasizing a constructionist approach using contemporary tools to create artifacts. The results underscore the significance of an appropriate rubric that should assess the appropriateness of the response, regardless of whether it was generated manually or through AI assistance. The intellectual output of the student should be measured. Learning tasks should be designed to utilize AI as an intelligent assistant. Students should receive direct instruction for foundational learning, followed by exploratory tasks to assess their competency before engaging in elementary construction activities to solidify knowledge. The final production of artifacts serves as an indicator of students’ mastery of the elementary content. Future research should focus on a refined pedagogy shifting from detecting AI usage teaching and learning with AI.