In recent years, Artificial Intelligence (AI) and large language models, such as ChatGPT, have gained attention for their ability to address diverse problems. However, their integration into educational settings raises concerns about academic integrity. Students increasingly rely on AI not only for essay-type questions but also for programming assignments, which is the focus of this study. This reliance undermines the evaluation of students’ genuine understanding and skill development. Existing countermeasures often fail to effectively identify AI-generated programs. To explore this issue, we carried out a two-stage case study: first, an experiment to assess ChatGPT’s accuracy in generating program solutions, followed by a focus group of experienced programming teachers. The group identified practices, strategies, and policies to address AI’s impact on academic integrity and assessment accuracy, while exploring opportunities to integrate AI for improving teaching and learning. These insights provide a framework to uphold academic integrity while harnessing the potential benefits of AI in education.