思想政治教育人工智能大模型语料库构建的问题审视和实践路径  

Examination and Practical Path of Construction of Artificial Intelligence Large Language Model Corpus in Ideological and Political Education

作　　者：沈强 SHEN Qiang(School of Marxism,Wuhu Institute of Technology,Wuhu,Anhui 241003)

机构地区：[1]芜湖职业技术学院马克思主义学院,安徽芜湖241003

出　　处：《绍兴文理学院学报》2024年第10期37-47,共11页Journal of Shaoxing University

基　　金：安徽省职业与成人教育学会项目“高职思政课教学方法研究——对错误思潮批判考察为例”(Azcj2022220);芜湖职业技术学院项目“算法推荐嵌入高校思政课实践教学基地数字智能化的逻辑机理与实践路径研究”(wzyrw202227)。

摘　　要：ChatGPT类人工智能大模型重塑了传统的内容生产方式与信息传播模式,在数据、内容、伦理、意识等层面给思想政治教育带来了数据遮蔽与技术异化、内容虚构与信息疫情、隐私侵犯与主体弱化、意识幻象与话语失真等问题困境。语料数据的体量、质量、种类、模态对人工智能大模型的训练和效果有着至关重要的影响,大模型的数据训练依赖性、内容缩放局限性、伦理涌现工具性和意识对齐类人性是导致思想政治教育大模型现实困境的主要原因。应对人工智能大模型介入思想政治教育问题风险,需要加强数据源头治理,以技术为底层逻辑,从数据可控、内容可信、伦理可靠、意识可行四个维度,构建专业化、宽领域、多模态、高质量的思想政治教育人工智能大模型语料库,强化语料数据收集、特征提取、内容微调和反馈更新全过程管理,提升对思想政治教育人工智能大模型的监控和干预能力,确保人工智能大模型的安全可控性,为人工智能提供优质思想政治教育服务奠定基础。Artificial intelligence models such as ChatGPT have reshaped the traditional content production mode and information dissemination mode,and brought problems such as data masking and technological alienation,content fiction and information epidemic,privacy infringement and subject weakening,consciousness illusion and distortion of consciousness to ideological and political education at the levels of data,content,ethics,and consciousness.The volume,quality,type,and modality of corpus data have a crucial impact on the training and effect of the AI model,and the data training dependence of the large model,the limitation of content scaling,the instrumentality ofemergence of ethics and the humanism of consciousness alignment are the main reasons for the practical dilemma of the large model of ideological and political education.To deal with the risk of artificial intelligence large models intervening in ideological and political education,it is necessary to strengthen the governance of data sources,take technology as the underlying logic,and build a professional,wide-field,multimodal,and high-quality corpus of artificial intelligence large models of ideological and political education from the four dimensions of data controllability,credible content,ethical reliability,and feasible awareness,strengthen the management of the whole process of corpus data collection,feature extraction,content fine-tuning,and feedback update,improve the monitoring and intervention ability of artificial intelligence large models of ideological and political education,ensure the safety and controllability of large AI models,and lay the foundation for AI to provide high-quality ideological and political education services.

关 键 词：思想政治教育 人工智能 大模型 语料库 路径 

分 类 号：D442.62[政治法律—政治学]