Through authentic book reviews as a source of humanism, an interactive AI-based application, capable of generating customized and personified evaluation, is thus produced focusing on multiple dimensions. According to comparative analysis with existing GenAI-based evaluation results, this study demonstrates that by assigning a human-like “evaluator” role to artificial intelligence, the application-centered assessment model developed in this study creates translation evaluations that are aligned to human readers’ preferences, thereby exhibiting distinctive humanistic appeal compared to traditional models. This research aims to improve the translation quality of Internet literature in English and provide linguistic support for its international dissemination.