Programming for the Next Generation: Transitioning from Traditional Models to Large Language Models in Coding Education for Kids

In an increasingly tech-based world, the integration of programming education at a global scale plays a crucial role in equipping students for the future. Programming languages such as Scratch, Alice, and Blockly have been available for some time and have effectively used visual and intuitive interfaces in introducing beginners to basic concepts in coding. However, with the emergence of Large Language Models like GPT-4, the vision of programming education is to evolve into real-time, contextual support and personalized feedback that makes them understand even better, improves problem-solving skills, and increases engagement. The synergy between blocks and LLMs facilitates an accessible, interactive, and highly supportive environment for learning among children. In this paper, we critically analyze block-based programming and LLMs to determine which offers a more advanced and impactful approach to the programming education of children. This review illustrates how these technologies supplement and compete with each other, shaping the future landscape of programming education by equipping the next generation with critical technological skills. We highlight the ethical concerns that might arise for LLMs. We point out how LLMs may adopt several strategies to address and counter such ethical challenges.