出版社:Conselho Nacional de Pesquisa e Pós Graduação em Direito
摘要:Algorithmic discrimination is one of the risks of the more and more frequent use of artificial intelligence systems. The paper criticizes the neutrality of algorithms, demonstrating that, in some cases, the digital revolution we are witnessing amplifies forms of discrimination already present in society and in the market and create new ones. Faced with such a complicated scenario, a balanced approach is proposed between mistrust and too much emphasis on artificial intelligence applications. In particular, it is advised that we not to give up the transparency and control of algorithms, to prevent them from being allowed to do what men are not allowed to, becoming a shield for the responsibility of some people and a tool to reduce other people to mere “ cases” to be classified, normalized or excluded.
关键词:Automated data processing;Algorithmic decision;Non discrimination;Transparency and controllability of artificial intelligence systems.