TEACHING AND LEARNING CREATIVE CODING WITH CONVERSATIONAL AI

ABSTRACT

The significant breakthroughs in large language models (LLMs) in recent years have enabled the development of numerous conversational artificial intelligence (AI) tools, such as OpenAI ChatGPT and GitHub Copilot. These tools have transformed the landscape of computing education by offering personalized assistance, instant feedback, debugging suggestions, and code completion support. Through a case study based on extensive teaching experience, we explore the benefits and challenges of using conversational AI in creative coding, providing practical guidelines for both learners and instructors. The research categorizes students into beginner, intermediate, and advanced levels, analyzing their interactions with AI tools across nine dimensions including usage frequency, prompt engineering skills, and learning experiences. Our findings indicate that conversational AIs can substantially improve learning efficiency and engagement, and they also require careful course design and management to prevent over-reliance and develop critical problem-solving skills. Further, Our findings reveal that while advanced students effectively leverage AI tools to enhance their capabilities, beginners and intermediate learners often struggle, facing issues such as over-reliance on AI-generated code and difficulty in critical assessment. The study highlights both the potential benefits of AI tools in providing personalized, immediate feedback and their limitations, including the risk of hindering fundamental skill development and creativity. Based on these findings, we propose practical guidelines for learners and instructors of creative coding on effectively utilizing these tools. We aim to lower the barrier for creative coding practitioners and inspire further exploration in computing education, particularly within the art and design fields.