This study examines the differences between human and AI-generated translations of Kate Chopin's The Story of an Hour in terms of literary style and irony. This short story was specifically selected as the research object due to its highly compressed narrative structure and pervasive, subtle irony, which provide a rigorous benchmark for evaluating interpretive translation competence. Employing a qualitative comparative methodology, the study analyzes specific passages from the source text alongside Indonesian translations generated by a human translator and the ChatGPT AI system. The investigation primarily focuses on irony translation, the preservation of literary style, translation quality according to Nababan's framework, and translational shifts in the AI output. The findings reveal that human translation more successfully preserves implicit meaning, ambiguity, emotional complexity, and stylistic features such as figurative language and symbolism. Conversely, AI-generated translation frequently reduces subtle rhetorical effects and produces a more neutral narrative tone by prioritizing readability and clarity through explicitation. According to Nababan's framework, while human translation excels in maintaining pragmatic depth and achieving higher acceptability, AI translation exhibits superior readability. Furthermore, semantic neutralization and explicitation are identified as the two primary shifts in AI translation. These results imply that although AI can generate fluid and readable sentences, it still struggles to preserve the aesthetic and interpretive dimensions of literary works.