With the rapid advancements in artificial intelligence, which could lead to better translation quality than before, there is a question to be explored: Which demonstrates better translation quality—formally trained MTI student translators or AI translation engines? To address this question, the present study examines MTI student translators from three universities in Guangdong Province, China, alongside AI translation engines (ChatGPT and Claude). Utilizing the Multidimensional Quality Metrics (MQM) Core model, the translation quality of both groups was quantitatively analyzed based on three primary indicators: Accuracy, Linguistic Conventions, and Style. The results indicate that, in the absence of specialized prompts, AI systems demonstrate better performances in accuracy, while human translators excel in linguistic conventions. Both groups showed proficiency in maintaining stylistic integrity. MTI students need improvement in reducing mistranslation and omission, while AI should look more into improving natural language flow and contextual sensitivity. These findings provide practical insights into the training and development of MTI students.