Whether protecting privacy, anonymity or property, encourage sharing, or to guarantee the absence of bias, the ethical issues raised by the collection and use of data are both numerous and ancient. In addition, there are the deleterious effects of some hasty applications of predictive systems based on the use of machine learning techniques, such as predictive justice, dynamic pricing or the distribution of police forces. However, the approach to these questions varies over time and between cultures. In addition, certain moral prescriptions, such as impartiality in the collection or the principles of finality and proportionality of the CNIL, go against the epistemological postulates posed by the most inflamed zealots of the masses of data. This will involve presenting these ethical issues and comparing them with the fundamental epistemological questions raised by data science.