EN
The growth of the digital economy has resulted in unprecedented increase of the amount of data generated by mankind. This vast volume and variety of data is a significant source of knowledge that allows researchers to open new research fields and to analyse existing problems more precisely. It is particularly important for scientists employing machine learning techniques in their studies. From a legal standpoint, however, the data typically belongs to the entity that collected it. In practice, there can be entities such as the owners of social networks (e.g. Instagram – Meta), online services (e.g. YouTube – Google), or Internet of Things devices like fitness bands. This data is protected by private law, thus, the rules on re-use of public data cannot be applied to it. Therefore, access depends on the will of the data holder. This fact has several adverse effects on the development of science and society. Attempts have been made at the European Union level to create legal instruments that facilitate access to privately owned data for academic purposes. One such instrument is the Digital Services Act. This article presents the analysis of the regulation in terms of balancing the interests of scientists and data holders, as well as the practical problems that may arise from its application.