基于生成性人工智能的校园欺凌预防智能化治理体系研究

Research on an Intelligent Governance System for Campus Bullying Prevention Based on Generative Artificial Intelligence

摘要: 近年来，校园欺凌事件频发，对学生身心健康造成严重威胁。本文基于对保定市某区县27所中小学32,586名学生家长的调查数据，采用因子分析和回归分析方法，探讨了家庭教育、家风家训、法治意识、欺凌认识等因素对校园欺凌发生的影响机制。结果表明，家长对校园欺凌的重视程度、认识水平、家风家训、家庭教育和亲子沟通质量与校园欺凌发生率呈显著负相关；而家长的法治意识、遇欺凌时的引导方式以及学生学习成绩与校园欺凌发生率呈显著正相关。学校对欺凌的培训对降低欺凌发生率没有显著影响。本研究创新性地提出，生成式人工智能技术可应用于构建校园欺凌预防智能体系，通过包含教育局层、学校层、学生及家庭层的三层两阶段方式，实现欺凌知识的个性化普及、实时咨询服务和欺凌应急处理的及时协助。基于布朗芬布伦纳的生态系统理论，提出了个人、家庭、学校、社会多层面联动的智能化校园欺凌预防策略，为从源头上针对性地遏制校园欺凌提供理论和实践参考。

Abstract: In recent years, school bullying incidents have occurred frequently, posing a serious threat to students’ physical and mental health. Based on survey data from 32,586 parents of students in 27 primary and secondary schools in a district of Baoding City, this study employs factor analysis and regression analysis to investigate the impact mechanism of factors such as family education, family values, legal awareness, and bullying perception on the occurrence of school bullying. The results indicate that parents’ emphasis on school bullying, their level of understanding, family values, family education, and the quality of parent-child communication are significantly negatively correlated with the incidence of school bullying. In contrast, parents’ legal awareness, guidance methods when encountering bullying, and students’ academic performance are significantly positively correlated with the incidence of school bullying. School training on bullying has no significant effect on reducing its occurrence. This study innovatively proposes that generative artificial intelligence technology can be applied to construct an intelligent system for school bullying prevention through a three-layer, two-stage approach, including the education bureau layer, school layer, and student and family layer, to achieve personalized popularization of bullying knowledge, real-time consultation services, and timely assistance in bullying emergency handling. Based on Bronfenbrenner’s ecological systems theory, the study proposes a multi-level collaborative intelligent strategy for school bullying prevention at the individual, family, school, and societal levels, providing theoretical and practical references for targeted suppression of school bullying from the source.