Abstract
<title>Abstract</title> <p>Artificial intelligence (AI) has become a central topic in global media discourse following the emergence of ChatGPT, shaping public understanding and policy debates on emerging technologies. However, traditional close textual analysis is limited in its ability to capture patterns in large-scale media data. At the same time, China and German-speaking countries represent contrasting sociopolitical contexts, reflecting different approaches to AI development and governance and providing a valuable basis for cross-country comparison. This study addresses these challenges by proposing a mixed-methods framework to analyze AI discourse in Chinese and German media, based on a corpus of nearly 30,000 news articles focusing on the period before and after the emergence of ChatGPT (2018–2023). The approach combines paragraph-level sentiment analysis and sentiment-informed topic modeling to identify how attitudes and themes co-occur in media coverage before and after ChatGPT, alongside qualitative analysis informed by Critical Discourse Analysis to interpret these patterns within broader sociopolitical contexts. The findings reveal a clear divergence between the two contexts. Chinese media predominantly frames AI in positive terms, emphasizing technological progress and national development, whereas German media present a more ambivalent discourse that combines optimism with concern and critical scrutiny. These patterns reflect differing sociopolitical orientations, with Chinese discourse aligned with developmental priorities and German discourse emphasizing risk governance and regulatory considerations. By integrating computational and interpretive approaches, this study contributes to comparative research on AI discourse and offers insights into how media discourses shape the societal understanding of AI across different contexts.</p>