现代情报2025，Vol.45Issue(11)：167-177,11.DOI:10.3969/j.issn.1008-0821.2025.11.015
基于Zipf-Mandelbrot定律的人工智能生成内容离散分布规律探析
Analysis of the Discrete Distribution Patterns of AI-Generated Content Based on the Zipf-Mandelbrot Law
摘要
Abstract
[Purpose/Significance]The breakthrough of Generative AI has triggered a surge in Artificial Intelligence Generated-Content(AIGC),rapidly reshaping the environment of information resources.However,systematic and quanti-tative analyses of the characteristics of AIGC or its impact on the development and usage of information resources in the future remain insufficient.[Method/Process]This study applied the Zipf-Mandelbrot Law to analyze the discrete distribu-tion patterns of AIGC across content units.It also compared these patterns of AIGC with 15 types of natural human corpora to explore the distribution characteristics of AIGC and the behavioral patterns of generative artificial intelligence(GAI)in content creation.[Result/Conclusion]The study finds that the distribution patterns of AIGC align with the"least effort prin-ciple"in human communication,suggesting that GAI achieves a balance between diversity and uniformity in information production across various tasks.In the era of artificial intelligence,GAI would serve as a significant new source of informa-tion production.Additionally,integrating AIGC into the information resources system could address the insufficiency of personalized content creation.The article also reveals that the Zipf-Mandelbrot parameters vary across different large mo-dels,demonstrating the potential for evaluating the results of models'performance.By analyzing the discrete distribution patterns of AIGC based on the Zipf-Mandelbrot Law,the paper proposes a quantitative evaluation framework for AIGC applicable to the fields of information resources management and linguistics.关键词
人工智能生成内容/生成式人工智能/信息分布/齐夫定律/Zipf-Mandelbrot定律Key words
artificial intelligence generated-content/generative artificial intelligence/information distribution/zipf's law/zipf-mandelbrot's law分类
社会科学引用本文复制引用
朱禹,蔡凌晨,陆泳溶,张逸勤,叶继元..基于Zipf-Mandelbrot定律的人工智能生成内容离散分布规律探析[J].现代情报,2025,45(11):167-177,11.基金项目
国家社会科学基金重大项目"学术'全评价'视域下中国特色哲学社会科学评价体系建设研究"(项目编号:24&ZD323) （项目编号:24&ZD323）
江苏省研究生科研与实践创新计划项目"生成式AI驱动的哲学社会科学'全评价'模型构建与验证"(项目编号:KYCX25_0130). （项目编号:KYCX25_0130）
