This article discusses the hypothesis supported by a group of authors in the 80's, about the general requalification of the workforce as a result of industrial automation. I criticize three main points: a) the new characteristics of qualification under automation are mechanically identified with an enlargement and deepening of the knowledge required; b) the concepts that designate the new characteristics of qualification are poorly defined, and c) there are no methodological instruments that allow a rigorous analysis of the evolution of qualification. Afterwards, I develop an alternative approach based on empirical evidence supplied by the literature.