Technological Empowerment and Humanistic Persistence: Defining the Application Boundaries of Artificial Intelligence in Basic Education

Abstract

The precipitate integration of Generative Artificial Intelligence (GenAI) into K-12 education has precipitated a dichotomy of “technological solutionism” and “pedagogical anxiety.” While Large Language Models (LLMs) offer unprecedented opportunities for personalized learning, they simultaneously pose significant risks to the developmental trajectory of students. This paper argues that the deployment of AI in basic education must not be governed solely by efficiency but by strict pedagogical boundaries. By synthesizing recent empirical studies and international policy frameworks (2023–2025), this study delineates three critical boundaries: the Cognitive Boundary, focused on mitigating “cognitive offloading” and preserving critical thinking; the Ethical Boundary, addressing data privacy and algorithmic bias under the guidance of UNESCO frameworks; and the Relational Boundary, reaffirming the irreplaceable nature of human empathy and moral modeling. The paper concludes that AI must serve as a “scaffold” rather than a “surrogate,” proposing a framework of “Structured Cognitive Offloading” to ensure technology empowers rather than erodes human agency.