Understanding the Interplay Between Artificial Intelligence, Online Propaganda, and Public Opinion Control: Insights from Yunnan's Cyberspace

Abstract With the continuous advancement of artificial intelligence technology, the risk identification and control of online public opinion have become increasingly crucial in the field of online propaganda. In particular, for a complex online space, such as Yunnan Province, risk identification and control of online public opinion are particularly important in the field of online propaganda. To grasp the risks of online public opinion, this study proposes a multi-platform network risk identification method based on artificial intelligence technology, which combines a network public opinion map with an optimized Bayesian model. This method utilizes artificial intelligence algorithms to construct a causal graph to reveal the development trajectory of network events, and effectively identifies and warns potential risks in online propaganda through optimized Bayesian models. The results showed that in the multi-ethnic and multicultural intersection of Xishuangbanna, the optimized network public opinion risk model had an accuracy rate of nearly 95% on different datasets, with the lowest being 90% and the highest sensitivity being close to 96%. After applying the optimization model, the distribution of public opinion risks became more concentrated, the overall risk coefficient decreased, and the number of positive sentiment data points significantly increased, resulting in a more concentrated distribution. This indicates that this study has theoretically enriched the research content of online public opinion prediction and risk control. Moreover, in practice, this study provides strong technical support and application demonstration for network public opinion management in Yunnan Province and other multi-ethnic areas, which has important theoretical value and practical significance.