Introduction
As Artificial Intelligence (AI) evolves into a "general-purpose technology" comparable to electricity, the necessity for foundational AI literacy has become paramount. Current educational paradigms often focus on university-level professional training, leaving a significant gap in primary education. This paper addresses the challenge of designing an AI curriculum for elementary students that balances technical proficiency with ethical understanding. By employing a backward-design approach, the authors propose a structured framework that moves beyond simple coding to foster a deep understanding of the "why" and "how" of AI.
Methodology: The AI Literacy Framework
The proposed curriculum is built upon a three-pillar competency framework designed to cultivate comprehensive AI Literacy in 5th and 6th-grade students:
- AI Knowledge (K): Focuses on the theoretical underpinnings, including definitions of AI, search algorithms, reasoning, and the mechanics of machine learning.
- AI Skill (S): Emphasizes practical proficiency through the use of AI tools and the development of "Computational Thinking" using block-based programming environments like "Entry."
- AI Attitude (A): Encourages a critical perspective on the social impact of AI, emphasizing ethical considerations and the ability to collaborate effectively with automated systems.
To ensure academic rigor, the authors conducted a comparative analysis of ten prominent AI textbooks and the CS2013 guidelines. These complex concepts were then "translated" into age-appropriate pedagogical strategies, including unplugged activities (physical logic games) and simplified visual explanations of advanced topics like Convolutional Neural Networks (CNNs).
Curriculum Implementation
The curriculum consists of eight modules. A key strategy involves using hands-on activities to demystify how computers "see" and process information. For instance, students engage in contour detection activities using pixel grids to understand how features are extracted from images—a foundational concept in computer vision.
Key Insights and Pilot Results
A pilot study involving 60 elementary students in South Korea revealed several critical insights:
- Distinguishing AI from Algorithms: Initially, students struggled to differentiate between standard software and AI. The curriculum successfully shifted this understanding, teaching students that AI is defined by pattern recognition in data rather than static, hard-coded rules. This was reflected in a dramatic improvement in test scores regarding AI identification ($t=10.998, p < 0.01$).
- Confidence and Competency: Students reported significant gains in their perceived ability to build AI applications using block coding ($t=2.272, p < 0.05$).
- Critical Awareness: While students entered the program with a generally positive view of AI, the lessons on social impact successfully fostered a more nuanced and critical understanding of the technology's risks, such as data bias and surveillance concerns.
Conclusion
The study demonstrates that complex AI concepts can be effectively taught to elementary students when structured through a literacy-focused framework. While the pilot study focused on the introductory module, the significant improvements in student understanding of AI definitions and social impacts provide a promising foundation. Future work will focus on evaluating the remaining modules and scaling this framework into a comprehensive K-12 roadmap for AI education.
Summary updated