摘要
人工智能属于通用目的技术,不仅能够引发产业变革,而且能够催生新科学研究范式。随着以生成式人工智能为代表的通用人工智能的发展,人工智能的应用成为社会科学研究范式创新的前沿。数据的高维度、编程语言的强结构性和长程关联性、以参数形式表征的隐式知识库和预训练的结构重整带来新知识和模式发现,为社会科学研究创造出新的发展空间。同时,数据隐私、价值观和研究结论的“黑箱”化带来了人工智能在社会科学应用中的新挑战。如何把大模型的知识发现和基于研究者实践经验的探索式研究相结合,形成人机共生演进的知识生产方式,是社会科学研究中人工智能应用的方向。
Artificial Intelligence(Al)is a general-purpose technology,not only sparking off industrial transformation,but also ushering in a new paradigm in scientific research.With the development of general AI,represented by generative artificial intelligence,the application of AI has become the forefront of innovation in the paradigm of social science research.The high dimensionality of data,the strong structural nature and long-range correlations of programming languages,the implicitly existing knowledge repository in parameterized form,and the structural reorganization through pre-training bring novel insights and pattern discoveries,thereby creating new developmental dimensions for social science research.Concurrently,the application of AI in social science encounters a series of challenges arising from data intricacies,value perspectives and the"black box"nature of research findings.How to address the amalgamation of data mining with large models and exploratory research grounded in practitioners'experiences,and establish a knowledge production mechanism for symbiotic human-machine evolution is the direction of AI application in social science research.
作者
刘刚
李响
李沁倩
刘捷
LIU Gang;LI Xiang;LI Qinqian;LIU Jie
出处
《理论与现代化》
2024年第2期80-91,共12页
Theory and Modernization
基金
2023年中国工程院战略研究与咨询项目“场景创新驱动的人工智能技术升级和产业发展”,项目编号:2023-HY-11
教育部哲学社会科学实验室专项基金研究项目,项目编号:H0122702。
关键词
人工智能
大模型
社会科学
对齐问题
隐私安全
artificial intelligence
big models
social sciences
alignment issues
privacy security
