摘要
在全球化新形势和世界政治经济新变局背景下,算法已经日益成为经济社会发展的重要支点,随着ChatGPT的横空出世,算法应用越来越多地支配着人类的生活,算法黑箱、算法操控、算法共谋、算法偏见、算法歧视等风险也随之而来,这些风险严重影响社会稳定乃至国家安全。而对国内外算法风险研究的系统梳理能够对全球算法风险的形势进行研判,进而为应对全球算法风险治理难题提供中国智慧与构想。文章通过系统梳理国内外主要数据库915条核心文献,构建基于“研究领域—研究主题—治理工具—治理措施”的算法风险研究框架,并分析算法风险具有学科的交叉性、复杂的交织性、突出的人为性、泛化的不确定性等特征,最后从加强情报学学科对算法风险的研究、加强对人工智能算法可解释性的研究、加强具有实际推广价值的算法风险实证研究、加强算法应用向善和算法服务从善的研究、加强对全球算法风险治理中国智慧与构想的研究5个方面对算法风险难点问题进行展望。
In the context of new globalization and new political and economic changes in the world,algorithms have increasingly become an important pivot for economic and social development.With the emergence of ChatGPT,algorithm applications are increasingly dominating human life.Risks such as algorithmic bias and algorithmic discrimination also follow,which seriously affect social stability and even national security.The systematic review of algorithmic risk research at home and abroad can study and judge the global algorithmic risk situation,and then provide Chinese wisdom and ideas for dealing with global algorithmic risk governance problems.By systematically sorting out 915 core documents in major domestic and foreign databases,this research constructs an algorithmic risk research framework based on“research field-research topic-governance tools-governance measures”,and analyzes that algorithmic risk has interdisciplinary,complex interweaving,outstanding human nature,generalized uncertainty and other characteristics,and finally from strengthening the research on algorithm risk in informatics disciplines,strengthening the research on the interpretability of artificial intelligence algorithms,strengthening the empirical research on algorithm risk with practical promotion value.A forward-looking outlook on the difficult problems of algorithmic risk is made in five aspects:strengthening the research on algorithm application for goodness and algorithm service for goodness,and strengthening the research on Chinese wisdom and conception of global algorithmic risk governance.
出处
《情报学进展》
2024年第1期172-204,共33页
Advances in Information Science
基金
国家社会科学基金重大项目“面向数字化发展的公共数据开放利用体系与能力建设研究”(项目编号:21&ZD336)
国家社会科学基金一般项目“数智环境下情报分析算法风险治理路径研究”(项目编号:22BTQ064)的成果之一
关键词
算法风险
算法治理
文献述评
未来展望
科技向善
algorithmic risk
algorithmic governance
literature review
future directions
science and technology for good
