Algorithmic discrimination and the criteria for special protection of structurally discriminated groups: an analysis under the Brazilian General Data Protection Law and the European Union General Data Protection Regulation

Abstract

The phenomenon of algorithmic discrimination primarily affects structurally discriminated groups, reproducing and amplifying deep-rooted societal biases using biased data, low-quality training datasets, or datasets with low representativeness. Given the need to ensure special protection for these groups and considering that the Inter-American Court of Human Rights (IACHR) establishes standards that define parameters for such protection, the following question arises: to what extent do the General Data Protection Law (LGPD) and the European Union’s General Data Protection Regulation (GDPR) incorporate differentiated criteria or elements of special protection for these groups? To address this, the study will first provide a contextualisation of algorithmic discrimination as a reproducer and amplifier of structural discrimination. Subsequently, the standards established by the IACHR will serve as a basis for understanding the parameters of special protection for structurally discriminated groups. Finally, the presence or absence of special protection parameters for vulnerable groups in Brazilian and European data protection regulations will be analysed. The research will be conducted using the hypothetical-deductive approach; the procedural method will be analytical, and the research technique employed will be bibliographic. The findings indicate that the special protection standards developed by the IACHR are only incipiently addressed in the analysed data protection regulations, which offer more generalised protection focused on the type of data being processed rather than specifically on its use to ensure differentiated protection for certain groups in situations of vulnerability and/or victims of structural discrimination.