从ChatGPT到Sora:生成式人工智能发展应用的意识形态风险审视  

From ChatGPT to Sora:An examination of the ideological risks in the development and application of generative artificial intelligence

作　　者：董翼 DONG Yi(School of Marxism,Chongqing University of Posts and Telecommunications,Chongqing 400065,China)

机构地区：[1]重庆邮电大学马克思主义学院,重庆400065

出　　处：《重庆邮电大学学报(社会科学版)》2025年第2期85-93,共9页Journal of Chongqing University of Posts and Telecommunications(Social Science Edition)

基　　金：国家社科基金项目:ChatGPT对高校思想政治教育的挑战及应对策略(23BKS182);重庆市高校数字思政实验创新中心课题:“数字思政”的主要特征、作用机制及育人样态研究(2024SZSZ-01);重庆邮电大学教改项目:民法典教育融入“思想道德修养与法律基础”课程的实践路径研究(XJG21245);重庆市教委人文社科重点研究基地项目:习近平新时代中国特色社会主义思想世界观和方法论的整体性研究(23SKJD067)。

摘　　要：作为生成式人工智能的阶段性代表,从ChatGPT到Sora,实现了从文本生成到视频生成的发展、从自然语言交互到物理世界特定模拟的跃进,凸显了人工智能技术的加速迭代之势。生成式人工智能的发展应用必然对社会生产生活产生重塑性、变革性影响,前所未有的意识形态风险即是其一。从属性上看,生成式人工智能的技术本质并未改变,其依然扮演着合目的性的工具角色,并具有鲜明的意识认知垄断、主体观念形塑、文化内核渗透等意识形态属性。从风险生成上看,生成式人工智能发展应用的意识形态风险,具有颠覆性技术的“创造性破坏”所带来的风险生成逻辑,是资本与技术深度“合谋”的必然结果,呈现显性风险与隐性风险并存的生成样态,因数据投喂、权威解构和监管困境产生的意识形态安全风险、意识形态话语风险和意识形态治理风险是其突出表征。从风险治理上看,如何从DeepSeek掀起的国产大模型浪潮出发,建构应对以ChatGPT、Sora为代表的生成式人工智能意识形态风险的策略路径,具有重要现实意义。把握好人工智能与人类智能的本质关系,即始终明确技术创造发展的核心价值在于为人类谋福祉,人的主体性参与是人机互动的基本前提;把握好治理认知与治理策略的并进关系,即坚持“技术向善”理念、注重提升主体智能素养,确立“共建共治”原则、着力增强敏捷治理效能,反思“人机共生”走向、应对多态应用场景治理;把握好价值引领与风险规制的统一关系,即坚守价值引领、共识凝聚、“红线”维护的治理主线,坚决防范西方错误思潮借助技术的反向“渗透”,坚持制度与监管互补的法治治理,是因应生成式人工智能意识形态风险的可行之策。As a representative of generative artificial intelligence(AI),the development from ChatGPT to Sora marks a leap from text generation to video generation,and from natural language interaction to specific simulations of the physical world,highlighting the accelerating iteration of AI technology.The development and application of generative AI will inevitably reshape and transform social production and daily life,with unprecedented ideological risks being one of the consequences.In terms of attributes,the technological essence of generative AI remains unchanged,continuing to play the role of a purposeful tool with distinct ideological attributes,such as monopolizing cognitive awareness,shaping subjectivity,and penetrating cultural core values.Regarding the generation of risks,the ideological risks associated with the development and application of generative AI follow the risk-generation logic of“creative destruction”brought by disruptive technology,which is an inevitable result of the deep“collusion”between capital and technology.These risks present both visible and hidden risks,with prominent features such as ideological security risks,discourse risks,and governance risks arising from data feeding,deconstruction of authority,and regulatory dilemmas.From the perspective of risk governance,how to construct a strategic path to deal with the ideological risks of generative artificial intelligence represented by ChatGPT and Sora from the domestic large language model wave set off by DeepSeek is of great practical significance.In terms of risk governance,it is crucial to understand the essential relationship between artificial and human intelligence,affirming that the core value of technological development is to benefit humanity,with human subjectivity being the basic premise of human-machine interaction.Additionally,it is essential to maintain the alignment of governance cognition and strategies,emphasizing the“technology for good”principle,improving the subject’s intelligence literacy,and establishing t

关 键 词：生成式人工智能 意识形态风险 ChatGPT SORA DeepSeek 

分 类 号：D64[政治法律—政治学] TP18[政治法律—中外政治制度]