Abstract: This article explores the revolution of generative translation in the era of artificial intelligence (AI), analyzing both its theoretical foundations and practical implications. After defining generative translation within the context of generative AI, the author details how these technologies are transforming the role of the translator, who is now positioned as an expert in post-editing and the management of hybrid projects. Drawing on concrete examples, the article highlights the strengths—such as handling complex contexts and producing creative renderings—as well as the limitations, particularly regarding the management of cultural nuances and biases inherent in training data. The article also proposes a classification of generative translation models and illustrates the redefinition of the concept of equivalence, shifting from an absolute correspondence to a nuanced approach to semantic and stylistic similarity. Finally, the author questions and rethinks classic dichotomies in translation studies (theory versus practice, translatable versus untranslatable, art versus science, author versus translator, etc.) and revisits translation theories—particularly the interpretative theory of meaning and the skopos theory—in light of AI advances, thereby opening up new perspectives for a smarter translation studies discipline. Keywords: Translation studies, generative translation, artificial intelligence, translation models, post-editing, equivalence, dichotomies, interpretative theory of meaning, skopos theory.