Abstract
The integration of generative artificial intelligence (GenAI) tools into low-code robotics platforms is sharply changing how educational robotics is integrated in science, technology, engineering, and mathematics (STEM) education. This paper presents a combination of an empirical study and a theoretical positioning to examine the integration of GenAI into low-code robotics within STEM education. The main aim of the study is to examine how GenAI facilitates progress in learn-to-code from basic syntax mastery, such as block-based programming, to advanced system architecture design that integrates multi-agent coordination, sensor fusion, and the ethical deployment of robots in STEM disciplines. Following PRISMA 2020 guidelines, 42 studies conducted between 2022 and 2026 were reviewed from databases including Scopus, Web of Science, IEEE Xplore, and ACM Digital to establish the theoretical positioning. In the empirical study, three low-code-generated STEM robot designs and workflows were analysed deductively to provide a thick description of what is feasible. Findings from both the empirical study and the review suggest that GenAI tools can make programming more approachable for beginners by helping them get started with simple code, providing debugging support, and offering Personalised scaffolding. At the same time, these tools can help learners gradually build the confidence and skills needed to take on more complex tasks, like designing modular systems and improving their work through ongoing refinement. From the analysis of developed robots, although generative AI offers many benefits, several challenges remain when Gen AI tools are used for programming. These challenges include learners still over-relying on AI-generated responses; equity issues related to language and cultural biases that continue to surface in Gen AI responses. This gap clearly shows a need for a well-structured teaching framework that thoughtfully blends human input with Gen AI collaboration. The Implications of this review suggest the need for a conceptual pathway for GenAI-augmented low-code robotics curricula and identify directions for educators and researchers to implement and validate the proposed framework.