Elvira V. Talapina

The principle of gender equality is closely linked to the right to non-discrimination, which has come into national legislations from international law. The evolution of the principle of equality from formal to substantive has been influenced by the Anglo-Saxon approach to European Union law, which focuses on equal treatment. In general, a fairly uniform model of anti-discrimination legislation has emerged worldwide.

The digitalization process has brought with it new threats, and a number of countries have already identified a gender gap in access to technology. However, the greatest risks are posed by digital gender discrimination - direct or indirect discriminatory actions that are based on automatic decisions made by algorithms. Such decision-making cannot be described as technology-neutral, as the algorithm may reflect the prejudices of programmers. The discriminatory decisions made by algorithms will be consistent and systematic, which is much more dangerous than individual human decisions. The causes of digital discrimination lie in modelling and making predictive recommendations based on discriminatory data, and in training algorithms based on discriminatory data. To overcome this complex problem - from the legal point of view - it is necessary to enforce the rule of transparency of algorithms as well as the decisions they make. It is necessary to conduct an audit of the algorithms - a special evaluation of the algorithm for potential violations of human rights. Also, the possibility and procedure of using artificial intelligence for decision-making must be regulated by laws. Some artificial intelligence systems should be banned outright, while others should be strictly controlled. The Proposal for a European regulation on artificial intelligence amply demonstrates this.