The increase in frequency of multidrug-resistant bacteria worldwide is largely the result of the massive use of antibiotics in the second half of the 20th century. These relatively recent changes in human societies revealed the great evolutionary capacities of bacteria towards drug resistance. In this article, we hypothesize that the success of future antibacterial strategies lies in taking into account both these evolutionary processes and the way human activities influence them. Faced with the increasing prevalence of multidrug-resistant bacteria and the scarcity of new antibacterial chemical molecules, the use of bacteriophages is considered as a complementary and/or alternative therapy. After presenting the evolutionary capacities of bacteriophages and bacteria, we show how the development model currently envisaged (based on the classification of bacteriophages as medicinal products similar to antibacterial chemical molecules) ignores the evolutionary processes inherent in bacteriophage therapy. This categorization imposes to bacteriophage therapy a specific conception of what a treatment and a therapeutic scheme should be as well as its mode of production and prescription. We argue that a new development model is needed that would allow the use of therapeutic bacteriophages fully adapted (after in vitro ‘bacteriophage training’) to the aetiologic bacteria and/or aimed at rendering bacteria either avirulent or antibiotic-susceptible (‘bacteriophage steering’). To not repeat the mistakes made with antibiotics, we must now think about and learn from the ways in which the materialities of microbes (e.g. evolutionary capacities of both bacteriophages and bacteria) are intertwined with those of societies.< Réduire