Mitigating Political Bias in AI-Generated News: A Comprehensive Analysis of NLP and Algorithmic Strategies

Yiming Sun; Zhuoer Lin

doi:10.62051/ijsspa.v4n2.31

Authors

Yiming Sun
Zhuoer Lin

DOI:

https://doi.org/10.62051/ijsspa.v4n2.31

Keywords:

AIGC (Generative Artificial Intelligence), Political Bias, Natural Language Processing, Quantitative Analysis, Algorithm Design, Dataset

Abstract

The application of artificial intelligence technology in news production is becoming more and more widespread, and this paper investigates the problem of political bias in AI-generated news. Through the comprehensive use of natural language processing, machine learning and quantitative analysis, the news generated by different AI models at home and abroad are systematically detected and evaluated to quantify their political tendencies, and the political bias exhibited by different models in news generation and its specific forms are analysed. A set of solutions to reduce bias is ultimately proposed, with improvements in terms of dataset construction, algorithm design, transparency and accountability mechanisms, supervision and integration of multiple values, aiming to enhance the fairness and objectivity of AI-generated news.

Downloads

Download data is not yet available.

References

[1] Cai Zhiquan. Value Evidence, Technological Pitfalls, and Strategic Responses to Generative Artificial Intel ligence Enabling New Quality Productivity. Southwest Finance 1-14.

[2] Ye Geze. (2023) . Diverse Representations of Mainstream Ideological Risks in the Age of Artificial Intelligen Resolution. Journal of Theory (06) , 62-69 . doi:10 . 14110/j. cnki.cn-37-1059/d.2023 .06 .009.

[3] ZhangXin. . (2023) . Data Risks and Governance Paths of Generative Artificial Intelligence. Legal Science (Jou rnal of Northwest University of Politics and Law) (05) , 42-54 . doi:10 . 16290/j. cnki. 1674-5205 .2023 .05 .006.

[4] Zhang Linghan. (2019) . The rise of algorithmic power, alienation, and legal regulation. Legal Business Resea rch (04) , 63-75. doi:10 . 16390/j. cnki. issn1672-0393 .2019.04.006.

[5] Zihan Xiong, Chonghao Cheng, Jun Chen & Dabei Chen. 2011 A new genus and species of the genus Pse udourostyla (Coleoptera , Staphylinidae , Staphylininae) from China. (2024) . Security risks facing generative artificial intelligence . Communications Enterprise Management (06) , 54-57.

[6] Tang Zheng & Lin Zilu. (2023) . Generative artificial intelligence and journalism: empowerment, risks, and fore sight. Journalism & Writing ( 11) , 97-104.

[7] Hu Feng & Zhang Chao. Theoretical Review, Risk Revealing and Governance Path of Generative Artificial Intell igence - A Perspective of ChatGPT. Journal of Nanjing University of Posts and Telecommunications (Social Science Edition) 1 -11 . doi:10 . 14132/j. cnki. nysk.20240625.001.