Nowadays it is generally accepted that the use of racial categories in biomedicine may reinforce social differences among racial categories; however, it has been claimed that this is the price that must be paid in order to reduce the differences in risk of complex diseases among them. In this paper I will show that this utilitarian argument can be accepted only under the “genetic hypothesis” about the existence of genetic differences among races that cause the differences in risk. “Races” used in biomedicine are categories constructed on the basis of self-declarations: they are not taxonomic categories but categories depending on personal beliefs about group membership characterized by several social and cultural properties, like having a specific behaviour, sharing a specific environment, eating specific food, etc. Given that, such categories are characterised by strong cultural and social differences that should be considered in causal explanations of differences of risk of diseases: why should the genetic hypothesis be preferred to the “social hypothesis” invoking social differences among races as causes of the differences in risk? I will show that the current biological theories in genome-disease associations support the role of environment in causing predisposition to complex diseases. If the differences in risk of diseases among racial categories are more likely to be caused by social differences than by genetic differences, the utilitarian argument cannot be accepted, since to reinforce social differences would mean in fact to reinforce the causes of the differences in the risk of diseases.