Abstract: [Purpose/Significance] The purpose of this study is to provide an in-depth analysis of the widespread application of artificial intelligence (AI) technology in the field of government data governance and its far-reaching implications, with a particular focus on the core issue of algorithmic discrimination. With the rapid development of AI technology, it has demonstrated great potential in government decision support, public service optimization, and policy impact prediction, but it has also sparked extensive debate on issues such as algorithmic bias, privacy invasion, and fairness. Through systematic analysis, this study aims to reveal the potential risks of AI algorithms in government data governance, especially the causes and manifestations of algorithmic discrimination, and then it proposes effective solutions to protect citizens' legitimate rights and interests from being violated, and to maintain government credibility and social justice. [Method/Process] This study adopts the literature induction method to extensively collect domestic and international related data on the application of AI in government data governance, including academic papers, policy documents, and case studies. Through systematic review and in-depth analysis, we clarified the specific application scenarios of AI algorithms in government data governance and their role mechanisms. On this basis, this study further identified the key factors that led to algorithmic discrimination, including but not limited to the one-sidedness of data collection and processing, the subjective bias of the algorithm designers, and the influence of inherent social biases on the algorithms. It then explored the potential risks of algorithmic discrimination, including exacerbating social inequality, restricting civil rights, and undermining government credibility, and provided an in-depth analysis through a combination of theoretical modeling and case studies. [Results/Conclusions] The results of the study show that while the embedding of AI technology in government data governance has significantly improved the efficiency and accuracy of governance, it comes with a risk of algorithmic discrimination that cannot be ignored. To address this issue, this study proposed a series of targeted prevention and control measures, including clarifying the principle of algorithmic fairness, formulating industry norms and standards, improving the accountability mechanism and regulatory system, and optimizing the data collection and processing environment, so as to effectively curb the phenomenon of algorithmic discrimination while making full use of the advantages of AI technology, so that AI technology in government data governance can truly benefit the people, and promote social fairness and justice.

Key words: artificial intelligence (AI), data governance, algorithmic discrimination, risk prevention and control, government data, information cocoon