The rise of short videos and live streaming has transformed digital governance, enabling governments to engage with citizens in real time. This paper explores the integration of artificial intelligence (AI) in government live streaming, focusing on its role in enhancing public opinion monitoring and sentiment analysis. We analyze current practices and introduce innovative solutions that leverage AI for improved content dissemination and audience engagement. Using the SO-PMI (Semantic Orientation Pointwise Mutual Information) algorithm, we conduct sentiment analysis on audience comments to capture emotional nuances. Additionally, large language models are employed for efficient live data processing, enabling real-time comment handling and generating responsive communication tailored to audience sentiment. The methodology includes robust data collection from social media platforms, employing API connections to retrieve comments during broadcasts. Extensive preprocessing involves filtering irrelevant content, normalization, tokenization, and lemmatization. Sentiment analysis categorizes comments as positive, negative, or neutral, while real-time monitoring allows presenters to adapt their messaging based on audience sentiment. Our experimental results demonstrate the effectiveness of AI-driven methods in managing real-time comments, providing accurate emotional analyses, and facilitating prompt responses. This study illustrates AI’s potential to engage citizens more effectively and enhance governmental communication strategies, ultimately fostering transparency and responsiveness in the digital age.