Racismo algorítmico e ferramentas de policiamento preditivo: proteção de direitos fundamentais ou um retorno a Cesare Lombroso?

Abstract

This work investigates the relationship between artificial intelligence (AI), algorithmic racism, and digital colonialism, focusing on the analysis of predictive policing tools. The study demonstrates how algorithms trained with data derived from human experiences (such as police interactions) reflect stereotypes, reinforcing existing racism within criminal prosecution processes. Indeed, predictive policing systems (such as PredPol and Palantir, discussed throughout the thesis) exemplify these practices by basing their crime predictions on biased data, which can potentially increase surveillance and even state violence against historically marginalized groups. This study falls within the study area "Democracy, Citizenship, and Fundamental Rights" of the Graduate Program in Rights and Fundamental Guarantees at the Faculdade de Direito de Vitória. Its aim is to demonstrate that the adoption of these technologies by developing countries under a rhetoric neutrality can be seen as a reflection of digital colonialism and a return to the precepts of 19th-century Lombrosian criminology. By examining these issues through a predominantly qualitative and bibliographic methodology, the research also hypothesizes that the creation of non- violating tools respecting fundamental guarantees is feasible, particularly in the field of criminal prosecution, with the premise that AI can be developed and applied in a way that upholds human and fundamental rights. To support this hypothesis, the thesis concludes by presenting concrete AI development projects aligned with the concept of foundational human rights, emphasizing the need for a state regulatory framework that prevents the perpetuation of historical and cultural biases in technological systems