The current education system urgently needs innovative learning spaces to help students develop the ability to solve real-world problems and improve their adaptability in the complex era of new technologies such as Artificial Intelligence (AI). This study describes an AI-coding hackathon designed to cultivate an inclusive and innovative learning space where students collaborate to solve real-world community problems. This project is designed according to Knowledge Building (KB) pedagogies, combined with AI-assisted learning support tools and chatbots, to provide students with learning scaffolds to help them develop programming and problem-solving abilities. The three-day program attracted students of different age groups from low socioeconomic backgrounds. The program is structured to guide students through basic machine learning concepts, basic Python programming syntax, and developing projects and solutions using prompt engineering to create an AI-supported application (digital books, chatbots, simple trained Machine Learning datasets) as the final product. The project emphasizes student agency, encouraging students to identify community problems, choose AI tools independently, and continuously revise solutions based on expert and group mentors’ feedback. Data collection included focus group interviews, classroom observations, and pretest-posttest questionnaires to measure students’ changes in programming skills, design thinking abilities, learning motivation, and knowledge building abilities. This study proposes a new pedagogical application, Knowledge Building, for AI education to promote a more equitable and meaningful AI learning experience.