Key points are not available for this paper at this time.
Amid the rapid advancement of generative artificial intelligence (AI), public perceptions and emotional attitudes toward AI have become increasingly complex. Existing research has predominantly focused on micro-level aspects such as risk perception and technology acceptance, with limited attention to its connections with national identity and cultural sentiment. Adopting a nationalism perspective, this study employs computational social science methods to examine whether and how nationalistic expressions emerge in public discussions on the Chinese large language model DeepSeek on the social media platform Weibo. We developed a lexicon-based methodology for nationalism, structured around five dimensions: national pride, national revival, anti-foreign, techno-nationalism, and cultural nationalism. Leveraging large-scale text mining and the Analysis of Topic Model Networks (ANTMN), we identify two distinct discursive clusters named the Utility cluster and the Sociopolitical cluster, and further conducted a comparative analysis of how nationalism was discursively articulated within each cluster. The results show that discussions of DeepSeek prominently reflect nationalistic sentiment, with techno-nationalism emerging as the most salient dimension. Significant structural differences were observed between clusters in the ways nationalism is articulated. This study expands the theoretical scope of AI public opinion research, proposes a quantifiable framework for analyzing nationalism, and offers new empirical insights into the national symbolism and collective emotions embedded in contemporary AI technologies.
Li et al. (Fri,) studied this question.