Artifitical intelligence literacy is increasingly recognised as a vital competency for young learners in an AI-driven future; yet targeted educational programmes for senior primary students remain scarce. This study evaluated a machine learning course developed for Hong Kong primary schools, designed to introduce foundational concepts—including supervised and reinforcement learning methods and algorithms such as artificial neural networks and K-nearest neighbours—through hands-on activities with AlphAI robots. The intuitive interface and visualisation capabilities of AlphAI robots with its software used in this study are designed to uncover machine learning as not a "black box", offering an innovative approach to demystifying machine learning for young learners. This study reported that sixteen teachers from five schools provided insights through focus group interviews after participating in a teacher development workshop and implementing the 6–8-hour course with students aged 11-13 years. An analysis of the interview data revealed that the course provided a clear and effective pedagogical framework, guided by the Attention-Engagement-Error-Feedback-Reflection (AEER) model, which teachers found particularly useful for lesson planning and instruction. Teachers reported that student engagement exceeded expectations, with students showing curiosity about machine learning mechanisms and demonstrating the ability to critically reflect on and compare robot behaviours driven by different learning methods—an outcome that impressed teachers. Despite these successes, teachers highlighted the need for more time and efforts to become proficient with the AlphAI software and suggested incorporating additional real-world examples to simplify complex concepts such as overfitting and backpropagation. Furthermore, participants emphasised the importance of increased school support to allocate sufficient instructional time and the value of peer collaboration both within and across schools to enhance teaching practices and student outcomes. These findings highlight the significance of machine learning curricula that integrate structured pedagogy with experiential robotics-based learning to develop literacy in young students. This study offers valuable insights for educators and policymakers aiming to effectively implement machine learning education in primary school settings.

Keywords: AI literacy education, teacher development, machine learning, primary school, teachers’ perspective.