教育领域生成式人工智能应用的伦理风险管理框架研究  

Research on Ethical Risk Management Framework for Generative Artificial Intelligence Application in Education

作　　者：王佑镁[1] 王欣颖 柳晨晨 WANG Youmei;WANG Xinying;LIU Chenchen(Research Center for Big Data and Smart Education,Wenzhou University,Wenzhou Zhejiang 325035)

机构地区：[1]温州大学大数据与智慧教育研究中心,浙江温州325035

出　　处：《电化教育研究》2024年第10期28-34,42,共8页E-education Research

基　　金：国家社会科学基金2023年度教育学重大项目“新一代人工智能对教育的影响研究”(项目编号:VGA230012)。

摘　　要：生成式人工智能对内容生产与传播产生了革命性影响,教育领域进入了智能化与风险性并存的人机协同时代,教育人工智能应用伦理风险频繁发生,急需通过标准化的风险评估工具进行科学管控。文章从技术与教育两个向度出发,将教育领域生成式人工智能应用的伦理风险归纳为技术本体风险、教育数据风险、机器算法风险和教育应用风险四大类。借鉴ISO31000风险管理标准与应用指南,搭建“风险识别—风险分析—风险评估—风险应对”四级流程的伦理风险管理框架。构建风险源检查表,以风险矩阵为主要评估工具,在实践中统计得出各类伦理风险综合等级。教育领域生成式人工智能应用的伦理风险管理框架,将为教育教学有效实施伦理风险管理工作提供借鉴价值。Generative artificial intelligence has had a revolutionary impact on content production and dissemination,and education has stepped into an era of human-computer collaboration in which intelligence and risk coexist.The ethical risks of educational artificial intelligence application occur frequently,and there is an urgent need for scientific management through standardized risk assessment tools.Starting from technological and educational dimensions,this paper summarized the ethical risks of generative artificial intelligence application in education into four main categories:technological ontology risk,educational data risk,machine algorithm risk,and educational application risk.Drawing on the ISO31000 risk management standards and application guidelines,this paper built an ethical risk management framework with a four-level process of"risk identification-risk analysis-risk assessment-risk response".A checklist of risk sources was constructed,with the risk matrix as the main assessment tool to statistically conclude the comprehensive level of various ethical risks in practice.This framework can provide reference for the effective implementation of ethical risk management in teaching and education.

关 键 词：生成式人工智能 伦理风险 教育人工智能 风险管理 风险评估 

分 类 号：G434[文化科学—教育学]