Motivated by the urgent need to investigate the possibilities for re-positioning the human translator and his/her educator in the machine translation (MT) age, this article explores the dynamics of the human-machine dance in the translation classroom. The article discusses the results of a collaborative learning experiment which was conducted in an MA course. In the experiment, the students carried out the emergent professional tasks of MT post-editing, pre-editing and error annotation. The results showed that the experiment helped the majority of the students re-position themselves as future experts of MT-related tasks by developing task-specific self-efficacy beliefs and raising their awareness of their self-concept as human agents and of the added value they could create. The educators, on the other hand, gained insights into the challenges faced by the students in the experiment, during which the boundary between teaching and learning became blurred. The educators further witnessed, first-hand, the need to transform translation classrooms into platforms of collaborative learning, where all involved can work together to discover ways of creating human added value in emergent MT workflows, which also requires the re-positioning of the translator educator in the MT age.