想象有道德的AI新闻业:机器伦理启示与价值对齐路径  

Imagining an Ethical AI-Driven Journalism: Inspiration from Machine Ethics and Approaches of Value Alignment

作　　者：常江 王鸿坤 CHANG Jiang;WANG Hongkun(School of Media and Communication,Shenzhen University,Shenzhen 518060,China;Institute of Global Communication,Shenzhen University,Shenzhen 518060,China)

机构地区：[1]深圳大学传播学院,广东深圳518060 [2]深圳大学全球传播研究院,广东深圳518060

出　　处：《山西大学学报(哲学社会科学版)》2025年第1期79-87,共9页Journal of Shanxi University(Philosophy and Social Science Edition)

基　　金：国家社科基金重大项目“数字新闻学理论、方法与实践研究”(20&ZD318)。

摘　　要：生成式人工智能作为关键技术类行动者的崛起,使新闻业的自动化程度发生质的深化,呈现出从流程自动化向判断自动化演变的趋势,并制造了一系列传统新闻伦理难以解释和应对的道德困境。具体而言,生成式人工智能让新闻价值判断成为计算的结果,促进了新闻模态的感官转向,形成了对话性的新闻分发机制。在这一过程中,新闻实践原有的德性原则因人类行动者对自身专业裁决权的不断让渡而被持续消解。鉴于此,对新闻伦理体系的重构工作应准确把握新闻行动主体网络的复杂性,借鉴技术哲学有关机器伦理的思想资源,探索人工智能新闻业价值对齐的实现方案。人工智能时代新闻伦理体系的建构应同时遵循两条路径:自上而下设置道德决策系统以矫正常规性伦理问题,自下而上建立反馈流程以辅助经验性伦理决策。As the key technological actor, generative artificial intelligence(generative AI) has qualitively deepened the degree of automation in journalism, exhibiting a trend of evolving from process automation to judgement automation, and resultsing in a series of moral dilemmas that traditional news ethics struggle to interpret and address. Specifically, generative AI has made news value judgments become the computational outcomes, facilitating a “sensory turn” in news modalitiy, and forming a conversational news distribution mechanism. During this process, human actors in journalism are continuously relinquishing their rights of making professional judgments, which has been increasingly undermining the traditional ethics of journalism. Given this trend, the reconstruction of the ethics of journalism should grasp the complexity of the social actor network in current news production practices, drawing on the intellectual resources of technology philosophy on machine ethics to explore how to implement value alignment in AI-driven journalism.The construction of the ethics of journalism should simultaneously follow two approaches: the top-down approach that sets up moral decision-making system to correct routine ethical problems;the bottom-up approach that establishes a feedback process to assist empirical ethical decision-making.

关 键 词：人工智能 新闻伦理 数字新闻 机器伦理 价值对齐 

分 类 号：G210.7[文化科学—新闻学]