Encouraging Student Use of AI in CS1: Rethinking Content, Delivery and Assessment
It has been well documented that generative AI tools are capable of solving traditional CS1-level problems as effectively as beginner students. Given this advancement, it is necessary to reconsider expectations for CS1 content as well as the pedagogical approaches we choose to adopt. This poster documents the experience teaching a recent course that encouraged students to use a generative AI tool while revisiting four pedagogical tools: Problem-Centric Instruction, Flipped Instruction, Competency Demonstrations and In- Person Code reviews. The course adopted a problem-centric approach; in a given unit, students were provided with a set of problems designed to highlight novel concepts, language features or problem styles. Borrowing inspiration from a flipped-style classroom, students were asked to prepare for class by developing solutions for these problems with the aid of an AI tool and then discuss the variations, similarities and effectiveness of the AI-based solutions during class sessions. The adoption of AI also necessitated modifications to traditional assessment targets and methodologies. Instead of evaluating the programs produced by the student (or the AI tool), competency demonstrations and one-on-one code reviews invited conversations with the students to better gauge their understanding. As a disruptive technology, AI can serve as a catalyst to revisit the overall objectives of introductory computer programming coursework and rethink what it means for students to engage with this field. It also provides an opportunity to revisit some of the pedagogical advancements that we should have been paying attention to all along.
