Against the backdrop of rapid advancements in AI translation technology, this article focuses on the highly challenging domain of translating classical Chinese poetry into English. Using Su Shi’s poetry as a specific case study, it systematically explores the innovative capabilities and inherent limitations demonstrated by AI translation. The research indicates that AI translation, leveraging its efficient processing capacity, consistent terminology management, big data pattern recognition, and multimodal generation potential, has brought significant process optimization and tool-based innovation to classical poetry translation. However, in-depth analysis grounded in cultural translation theory and hermeneutics reveals structural deficiencies in AI’s ability to poetically reconstruct artistic conception, deeply interpret cultural connotations, flexibly transform linguistic artistry, and particularly in executing “creative rebellion”—a core competency of literary translation. Through comparative analysis, this paper argues that relying solely on AI cannot independently achieve high-quality poetry translation. Ultimately, the study proposes establishing a collaborative workflow model centered on human translators with AI serving as deep-level assistance. By leveraging complementary strengths between humans and machines, this approach aims to achieve dual enhancements in efficiency and artistic quality for classical Chinese poetry translation into English, offering a practical pathway for the effective and profound dissemination of Chinese culture abroad.