The aim of this paper is to present the basic ethical dilemmas in the creation and functioning of visual databases in the social sciences. The author starts from the assumption that social research is a form of power and therefore requires regulation in regard to its methods and the management of the collected empirical material. Three main dimensions of the ethics of visual research are presented: in-formed consent; anonymity and confidentiality; and presentation, dissemination and anonymization. The author discusses these dimensions in regard to proposals for the creation of open databases under the Horizon 2020 Open Research Data project and then presents his own proposals concerning matters of importance in creating visual databases.