Tomada de decisões algorítmicas: uma análise da inteligência artificial nas decisões judiciais pela perspectiva da ciência da informação

Abstract

Data evolves to become valuable information every day. Collected data needs safeguards, especially in the face of technological innovations, often being organized and used to establish standards that impact decision-making procedures. Artificial Intelligence (AI) is part of the reality of individuals every day, in the Judiciary and in different areas. Faced with its use, many concerns arise, such as, for example, the need for transparency, security, ethical and legal issues, the search for a safe and reliable AI that also provides responsibility for the use of information. In this context, the objective of this dissertation is to analyze the impact of automated algorithmic decision-making considering the Information Science perspective. Among the methodological aspects, a bibliographic and documentary survey was adopted, analysis between national and foreign legislation, AI regulations and bills, uses and contexts associated with algorithms that establish guidelines for their transparency and ethical uses in decisions were adopted. of decisions. Regarding the approach, this research adopts qualitative and quantitative aspects. In relation to the research objectives, it is classified as correlational. The theoretical framework is based on two thematic axes. In the first, questions are presented about the use of AI for judicial algorithmic decision-making, presenting the implementations and how the discussions about AI regulation and transparency in the use of algorithms are going, evaluating the implications of automated decision-making, understanding the entire context of the information. In the second, some concepts and characteristics are presented that are associated with the increasing use of new technologies with the social paradigm of information science, the need for transparency in access and use of data. The results reveal that there is local and international concern with the implementation of safe and reliable systems, which are in progress and approved in some countries, and that there is the presence of legal provisions on algorithms and possible tracking of automated decisions. Finally, these results showed that the use of AI in automated decisions transformed the workflow with efficiency and positive results and that understanding data processing is interesting for information quality and security, aiming at the use of information.