生成式人工智能教育应用的安全风险及应对策略  

Security Risks and Countermeasures of Generative AI Educational Applications

作　　者：兰丽娜 黄莎莎 石瑞生 Lina LAN;Shasha HUANG;Ruisheng SHI(School of Humanities,Beijing University of Posts and Telecommunications,Beijing 100876;School of Cyberspace Security,Beijing University of Posts and Telecommunications,Beijing 100876)

机构地区：[1]北京邮电大学人文学院,北京100876 [2]北京邮电大学网络空间安全学院,北京100876

出　　处：《中国教育信息化》2025年第4期46-56,共11页Chinese Journal of ICT in Education

基　　金：中国高等教育学会2023年度高等教育科学研究规划课题“生成式AI支持下的高等教育数字化推进策略与实践研究”(编号:23PXZ0419);2023年度中国—中东欧国家高校联合教育项目“人工智能时代生成式AI促进高等教育数字化变革研究”(编号:2023293)。

摘　　要：以ChatGPT为代表的生成式人工智能一经推出就在社会各界引起火爆反响,其强大能力预示着新的通用人工智能时代的到来,必将对各行各业产生深远影响。但相关安全问题引发人们对于其应用的担忧。对于以立德树人为根本任务的教育行业,如何在教育教学中安全、有效地使用生成式人工智能,是相关研究必须回答的关键问题。由此,从教育、技术、伦理多个角度分析生成式人工智能教育应用过程中的安全风险问题,并提出应对策略。其中,安全风险主要包括:生成式人工智能的不当应用,可能造成学生学习自主性降低、被错误信息误导等负面影响风险;生成式人工智能的数据安全和隐私保护不足,可能造成教育数据隐私泄露风险;生成式人工智能教育应用缺少伦理约束,可能带来伦理安全风险。据此,从教育理论、安全技术、科技伦理三方面提出对应的应对策略,主要包括:以培养学生批判性思维为目标的生成式人工智能与教育融合路径;面向生成式人工智能教育数据处理全生命周期的数据安全与隐私保护策略;生成式人工智能教育应用的科技伦理治理理念和方法。The emergence of generative artificial intelligence(GAI),represented by ChatGPT,has sparked widespread attention across society,with its formidable capabilities heralding a new era of general artificial intelligence that is poised to profoundly impact all sectors.However,its associated security concerns have raised apprehensions regarding its applications.For the education sector,whose fundamental mission is fostering virtue and nurturing talent,the critical question that must be addressed is how to safely and effectively integrate GAI into teaching and learning practices.This paper analyzes the security risks in the process of generative AI education application from the perspectives of education,technology and ethics,and puts forward corresponding countermeasures.Security risks include:(1)the improper use of GAI may negatively impact students’learning autonomy and expose them to misinformation;(2)inadequate data security and privacy protections could lead to breaches of sensitive educational data;and(3)the lack of ethical constraints in GAI applications may result in ethical safety hazards.In view of the above risks,this paper puts forward corresponding countermeasures from three aspects of education theory,safety technology and science and technology ethics,including:the integration path of generative AI and education aimed at cultivating students’critical thinking;Data security and privacy protection strategies for the full life cycle of generative AI educational data processing;The scientific and technological ethical governance concepts and methods of generative AI educational application.These strategies are expected to provide references for the research and practice of generative AI educational application and promote the healthy development of the educational application of AI technology.

关 键 词：生成式人工智能 ChatGPT 人工智能教育 数据安全 隐私保护 

分 类 号：G434[文化科学—教育学]