ChatGPT的传播特质、逻辑、范式  

Research on the Communication Characteristics,Logic and Paradigm of ChatGPT

作　　者：王建磊[1,2] 曹卉萌 WANG Jian-lei;CAO Hui-meng(School of Communication,Shenzhen University,Shenzhen,Guangdong,518060;Sino-foreign Cultural ExchangeCenter of the Ministry of Education,Omnimedia Industry-Education Integration Base(Shenzhen University),Shenzhen,Guangdong,518060)

机构地区：[1]深圳大学传播学院,广东深圳518060 [2]教育部中外人文交流中心“全媒体产教融合基地(深圳大学)”,广东深圳518060

出　　处：《深圳大学学报（人文社会科学版）》2023年第2期144-152,共9页Journal of Shenzhen University：Humanities & Social Sciences

基　　金：国家哲学社会科学后期资助项目“智能推荐型视频产业价值链研究”(21FXWB003);广东省教育科学“十三五”规划研究项目“粤港澳大湾区青年创新与文化融合研究”(2020GXJK077)。

摘　　要：在技术路径上,ChatGPT通过前所未有的大语言模型训练和对人类反馈方式的学习,具备了生成对话的能力,在语词重组、自我表达修正和信息整合方面的强化使其成为知识生产的新引擎,又通过掌握特殊标记和经验性学习能给出一定的情感反应,从而形成了对人类的“超模拟”。人们在索取答案、发起挑战、提出问题的同时形成自发投喂AI的群体行为,ChatGPT据此反向延伸人类,不断获得更好的连接能力和嵌入社会运转系统的机会,其生长速度显现出新的摩尔效应;反向的技术依赖、动态的连接与嵌入、加速堆叠硬件和升级算法等构成其传播逻辑。最后,ChatGPT开启了“人机共在”的新主体时代,人机协同将作为长期的社会解决方案,而针对人机交互的复杂性和AIGC的衍生问题,也可能陷入诠释经验的多重解释的迷乱之中,故需要在突破“主客二元论”和“诠释经验主义”两个传统研究范式的基础上进行价值重构。Under the investigation of the technical path,ChatGPT has the ability to generate dialogue through the unprecedented training of large language model and the learning of human feedback mode.The strengthening of its ability in word restructuring,self-expression correction and information integration has made it a new engine of knowledge production,and it can give certain emotional response through mastering special marks and empirical learning,the above elements forming a"super simulation"of human beings;At the same time,when people ask for answers,they form a group behavior of feeding AI spontaneously,therefore ChatGPT extends human beings in the reverse direction.With better connectivity and the opportunity to embed into the social operating system,ChatGPT brings new Moore Effect.It can be said the reverse technology dependence,dynamic connectivity and embedding,.accelerating stacking hardware and upgrading algorithms become ChatGPT's propagation logic.Finally,ChatGPT has opened a new subject era of"human-computer coexistence".Human-computer collaboration will be a long-term social solution.However,the complexity of human-computer interaction and the derivative problems of AIGC may also fall into the confusion of multiple interpretations of interpretation experience.Therefore,it is necessary to conduct value reconstruction on the basis of breaking through the two traditional research paradigms of"subject-object dualism"and"interpretation empiricism".

关 键 词：ChatGPT 传播特质 传播逻辑 传播图景 认知范式 人工智能 基础设施 

分 类 号：G206.2[文化科学—传播学] G206.3