This paper explores the linguistic and rhetorical means used to express emotionality in Bosnian- -Herzegovinian media discourse on artificial intelligence (AI). The corpus consists of 15 newspaper articles published in December 2024 on the portals Media.ba, Klix.ba, Vijesti.ba, Avaz, and Oslobođenje. The analysis combines quantitative and qualitative methods within the DIMEAN discourse-linguistic model, focusing on the intratextual level (lexical-stylistic and semantic analysis). The results show the presence of both explicit (directly named emotions and value judgments) and implicitemotional expressions (metaphors, personification, hyperbole, semantic contrasts), with the latter being more frequent. The public discourse on AI oscillates between emphasizing potential benefits (revolutionary improvements to human activities) and concerns (job losses, unpredictable consequences). References to authoritative figures (Bill Gates, Geoffrey Hinton, Lisa Kudrow) further intensify the emotional dimension, as these public figures’ attitudes resonate with readers. Overall, the analysis indicates that Bosnian-Herzegovinian media, by portraying AI as both advantageous and risky, significantly shape public knowledge and attitudes toward this technology. Future research could include a larger sample and a longer time span to examine more thoroughly the evolving nature of AI discourse.