The application of Artificial Intelligence (AI) within translation studies has become a focal point of interest, especially concerning the translation of literary works as they represent almost unsolvable tasks. This research explores the complexities and strategies related to evaluating translations produced by AI, with a particular emphasis on Shakespeare's Sonnet 7 and Eliot's Middlemarch. The central goal is to create a communicative framework that assesses both the fidelity of the conveyed meaning and the retention of literary qualities and cultural subtleties. An analytical approach is employed, utilizing AI translation tools such as DeepL and Google Translate for comparative analysis. This research assumes that successful literary translation encompasses both semantic fidelity and stylistic resonance, reflecting the intricacies of the source material. The findings reveal that while AI-generated translations often convey the semantic meaning and the overarching themes, they frequently struggle with preserving poetic devices, rhythm, and tonal subtleties which results in hindering the original author‘s aesthetic message. For instance, the assessment of Shakespeare's Sonnet 7 indicates a partial effectiveness in capturing connotative meanings, while Eliot‘s Middlemarch shows a more favorable preservation of narrative tone. Ultimately, this paper posits that a detailed communicative framework based on linguistic elements is essential for critically evaluating AI translations in literature, recommending further refinement of AI tools to enhance their sensitivity to aesthetic dimensions and reader engagement.