This chapter intends to analyse the tensions between access to data for training AI systems and the control of content by intellectual property rights owners, framed in the European Data Strategy. This overlap focuses on the use of protected works as input for training, regulated by the Text and Data Mining (TDM) exception of the Copyright Directive. The application of this exception is only legitimate if it complies with the so-called Three-Step Test and the rightholders’ right to expressly reserve the use of their works (‘opt-out’). In addition, the AI Act requires providers of generative AI systems to implement appropriate policies to respect copyright by identifying and observing such reservations of rights, including through advanced technologies.