This paper focuses on the narratives surrounding the birth of AI technologies in the 50s and how they relate to translation, then and now. Geographically centered on the USA, research into automated translation was spurred on by the Cold War in the 1950s, as it has been of late, by the “war on terror”. On a philosophical level, automated language processing is rooted in cybernetics, theorised by Norbert Wiener as the science of command and control. In parallel, the paper investigates current narratives on AI research, as they are spun by the Californian tech entrepreneurs : the notion of progress, the sense of a mission, the equation between machine and brain. The related narratives of anthropomorphism, of the moonshot ideology, of algorithmic objectivity, also need to be questioned and situated : all play into a scenario of unprecedented command and control over human exchanges.