The present study is a comparative approach to the human translation (HT) quality produced by 22 students in the 3rd year from the Department of Translation, Interpretation and Applied Linguistics, Moldova State University, and the pure machine translation (MT) of the same text by Systran, Google Translate and DeepL. The source-text proposed for translation is a selection of three sentences excerpted from a prescription in English. The research questions are formulated as follows: What types of errors are made by HT and MT? How does the quality of the two types of translation impact the final product of translation? Is MT output of lower quality compared to HT? Comparative analysis and qualitative analysis are the main research methods used in the study. The results show that although both humans and machines are prone to making errors, they appear in different ways and at different rates in both cases. The study also raises awareness to the fact that even if MT has achieved good improvements since its initial stages, it can achieve its highest potential only if accompanied by human post-editing. Prospective studies that would derive from this research will be oriented towards undergraduates’ understanding of the role and importance of human post-editing of MT for the English-Romanian language pair, the contexts and the types of texts tolerated by MT.