¿Son fiables las decisiones de las Administraciones públicas adoptadas por algoritmos?

Abstract

Public administrations are already using artificial intelligence in decision-making as support to public officials but also to automate decision-making. Despite the benefits that this may entail, throughout this article it becomes clear that the use of artificial intelligence in public decision-making cannot be done in any way since this can generate several problems such as opacity, legal uncertainty, errors, biases or breaches of personal data protection. The mechanisms already provided by law are not enough to avoid these problems as they have not been designed to face the use of artificial intelligence in public administrations. Therefore, the article concludes that public administration should strengthen the reliability of algorithms through the guarantee of their legality, security and ability to obtain quality public decisions.