A tardia responsabilidade civil na discriminação algorítmica e a regulação estatal para sistemas de governança e compliance

Abstract

Algorithm systems are increasingly present and interfere in people's lives. The research problem is the harmful capacity of algorithmic systems, on a large scale and without borders, especially when they produce or encourage discrimination, without the respective repair. The hypothesis is that civil liability is not efficient to generate incentives to prevent damage, nor to repair it properly. As a general objective, it will be demonstrated that, in addition to state regulation, it is necessary to effectively demand governance and compliance structures, with transparency, risk management and accountability regarding the algorithmic configuration and its criteria, allowing its use as a tool of trust. As specific objectives, the algorithms will be discussed, conceptualizing them to understand their formation and highlight the existing legislation; the problem will be analyzed, considering the impacts in reality; Finally, it will work on the hypothesis of governance and transparency as a tool for compliance in the use of algorithms. The methodology is inductive, making a bibliographic review, seeking to discuss real occurrences eventually reported by the media. The findings reveal that the State has already been reacting to algorithmic discrimination, establishing legal requirements and imposing millionaire fines on large corporations, but it demands cooperation from data authorities around the world, as well as companies within the scope of innovation systems.