Browsing by Author "Junkkari, Jonni"

Article 22 of the EU General Data Protection Regulation 2016/679 (GDPR)arguably forbids the data controller from using solely automated processing to make decisions which produce legal effects concerning the data subject or similarly significantly affects the subject. It can be argued that the legislature assumes that solely automated processing of decisions is disadvantageous for the data subject’s legitimate interests. The purpose of this thesis is to investigate the hypothesis that solely automated processing of decisions is in some cases advantageous for the data subject’s legitimate interests, especially economic ones, and human involvement at times disadvantageous. The focus of the research is on situations where the data controller is a private enterprise, and the data subject is a natural person in a customer relationship with the enterprise. The economic impact that human involvement in automated decision-making may have on data subjects will be demonstrated through a thought experiment. In addition, an evaluation framework for determining the “goodness” of human vs algorithm decisions will be established. The scholarly debate surrounding GDPR Article 22 will be discussed and the argument that the subject’s economic interests should be part of the legal consideration for GDPR Article 22 will be established in the literature review. This will be complimented with a comparison of how certain other EU and non-EU legislation have approached legislating automated decision-making, with a particular interest on human involvement. Analysis of available case law connected to GDPR Article 22 will be analyzed for the protection it provides to data subject. The potential negative economic effect for data subjects of human involvement or intervention in automated decision-making will be established based on established economic concepts. The nature of human and algorithmic decision-making will be discussed, reaching the conclusion that the core problem of advanced algorithmic automated decision-making is not caused by lack of human involvement in decision-making but lack of transparency and explicability of the algorithm.