Veštačka inteligencija u pravnom sistemu EU
Artificial intelligence in the EU legal system

Author(s): Dragan Prlja, Gordana Gasmi, Vanja Korać
Subject(s): Law, Constitution, Jurisprudence, Constitutional Law, International Law, ICT Information and Communications Technologies, EU-Legislation
Published by: Institut za uporedno pravo
Keywords: EU; Artificial Intelligence Regulation; AI
Summary/Abstract: The adoption of the EU Artificial Intelligence Regulation will represent a significant step towards creating a legal framework for the functioning of artificial intelligence systems, both in the EU and throughout the world. In this way, trust in artificial intelligence systems will be built, the rights of individuals will be protected, the development of reliable and safe artificial intelligence systems will be enabled, legal certainty will be created and the market for services and products based on artificial intelligence technology will develop. The development of new technologies has created potential dangers for individuals and companies, but quality legal regulation and its application in practice, along with an effective system of sanctioning those who do not respect legal norms, is an adequate response that leads to the creation of a better organized and safe society in the future. The level of adverse impact of the artificial intelligence system on the fundamental rights protected by the EU Charter on Fundamental Human Rights is particularly important when classifying the artificial intelligence system as high-risk. These rights include the right to human dignity, respect for private and family life, protection of personal data, freedom of expression and information, freedom of assembly and association and non-discrimination, consumer protection, workers' rights, rights of persons with disabilities, the right to an effective legal remedy and to a fair trial, the right to defense and the presumption of innocence, and the right to good governance. Technical inaccuracies of artificial intelligence systems intended for remote biometric identification of individuals can cause biased results and have discriminatory effects. This is especially important when it comes to age, ethnicity, gender or disability. Hence, real-time remote biometric identification systems and post-remote biometric identification systems should be classified as high-risk, subject to special requirements regarding event logging and human control functions. Artificial intelligence systems used to assess the credit rating or creditworthiness of individuals should be classified as high-risk artificial intelligence systems, as they determine the access of those individuals to financial resources or basic services such as housing, electricity and telecommunications services. Artificial intelligence systems used for this purpose may cause discrimination against persons or groups, for example on the basis of racial or ethnic origin, disability, age, sexual orientation, or may create new forms of discrimination.

• Details
• Contents