Enhancing primary school students' Chinese reading ability through artificial intelligence at Huacheng foreign language school, Sichuan province

Abstract

To address the allocation imbalance of reading teaching resources in western China's primary schools, this study took the Doubao AI platform as an auxiliary teaching tool, selected 50 sixth-grade students from a provincial AI demonstration primary school in Panzhihua City, Sichuan Province, as subjects, and conducted an 8-week single-group pre-test-post-test reading intervention covering pre-class, in-class, and post-class phases. Following the 2022 Compulsory Education Chinese Curriculum Standards, the research focused on three core reading dimensions: detail extraction, central idea summarization, and logical analysis. Data were collected through standardized reading tests, student cognitive questionnaires and classroom teaching logs, and analyzed by paired-sample t-tests, reliability and validity analyses, and descriptive statistics. Findings show that after AI-assisted reading intervention, participants achieved statistically significant improvements (p<0.01) in the three core dimensions and overall reading scores, confirming the intervention's effectiveness. The proportion of students with intermediate to advanced reading proficiency increased from 24% to 82%, and all lowest D-level students were eliminated. Students generally gave positive evaluations of the platform, yet the tool has clear limitations in cultivating higher-order critical thinking. This study integrates constructivist theory, cognitive load theory, and the Technology Acceptance Model (TAM) to explain how Large Language Models (LLMs) improve primary school reading instruction, proposes targeted teaching strategies for relevant educational stakeholders, and provides localized empirical evidence for promoting AI reading tools in rural primary schools of western China.