Police violence is a daily reality in Brazil, especially in marginalized areas such as Maré, one of the largest favela complexes in Rio de Janeiro. Over 140,000 Maré residents live under constant risk due to large-scale police operations. These actions, justified by the war on drugs, often result in systematic human rights violations, including physical violence, property damage, home invasions, and extrajudicial killings.
In response to the persistent underreporting of human rights violations, the Data & Justice project was formed as an independent initiative that leverages artificial intelligence (AI) and data analysis to map and classify these violations. The initiative employs natural language processing (NLP) to automate the classification of reports, a process that currently relies on manual reviews conducted by individuals who categorize reports based on subjective interpretations—a method that is slow, inconsistent, and prone to human error.
Maré as a lens into problems with policing and surveillance
In places like Maré, where large-scale police operations are a constant reality, residents frequently report violations they have suffered. However, these reports, often numerous and generated under distressing circumstances, are documented in fragmented, unstructured ways, making it challenging to detect broader patterns of abuse. Data & Justice seeks to change this by using NLP to quickly and accurately categorize these reports, enabling the identification of recurring patterns that might otherwise go unnoticed. By training algorithms on a set of data that has already been categorized, our model is able to rapidly classify new reports and uncover hidden connections within the data. These insights can then be used to support advocacy efforts, hold authorities accountable, and inform policy interventions aimed at protecting vulnerable populations.
The situation in Maré is not an isolated phenomenon but rather is part of a system of public security policies that affect marginalized communities worldwide. In many countries, repression strategies based on the war on drug enforcement have led to the militarization of police forces, excessive use of force, and the criminalization of vulnerable populations. At the same time, artificial intelligence and other technologies are increasingly being incorporated into these policies, often reinforcing the surveillance, oppression, and persecution of minorities. In contrast, Data & Justice could reverse this dynamic by leveraging AI to uncover patterns of abuse and offer data-driven alternatives to official discourses, promoting transparency and state accountability.
The use of AI and NLP to monitor human rights violations has the potential to transform how police violence is recorded and addressed. This approach enables the analysis of a large volume of information quickly and accurately, surpassing traditional methods that rely exclusively on manual review of reports. As an example, reports of human rights violations can be collected and organized in a spreadsheet or database, with each entry corresponding to a testimonial or incident description. NLP techniques can then be applied to process this unstructured information, allowing for the automatic classification of reports by type of violation, location, or responsible actor. Instead of relying solely on time-consuming and error-prone manual review, this approach enables large-scale analysis and the identification of systemic patterns that might otherwise remain obscured.
However, it is widely recognized that AI has been increasingly used by governments as a tool for surveillance and repression, often under the guise of public security. Technologies such as facial recognition and predictive analytics frequently deepen racial profiling practices and reinforce repressive policies.
Using AI tools for justice
In response, the Data & Justice initiative reverses this logic, employing AI to identify patterns of abuse, enhance transparency, and challenge official narratives, promoting state accountability and shedding light on injustices that are often ignored.
Furthermore, this methodology can have a real impact on public policy formulation and legal debate. A systematic approach to collecting evidence of human rights violations enables the production of evidence-based reports, which can support legal actions and put pressure on authorities to make structural changes. The methodology proposed could also reshape public policy and legal frameworks. Systematic data collection on rights violations has the potential to support evidence-based reports, legal actions, and pressure for structural change. For instance, the initiatives Fogo Cruzado and GENI/UFF already monitor armed operations in the community and show how detailed, disaggregated data can affect judicial rulings and international monitoring. This work has shed light on the role of the state in perpetrating lethal violence, particularly through police-led massacres—223 of the 305 massacres between 2016 and 2021 involved police operations, resulting in 878 deaths.
From this perspective, AI helps give voice to historically silenced communities, ensuring that their reports are not treated as isolated cases but as part of a systemic issue. As a result, the Data & Justice initiative not only documents abuses but also provides concrete tools to support advocacy actions and demand structural changes. By transforming scattered data into organized and accessible information, the project strengthens the right to life, security, and dignity while also challenging the impunity of the state and reinforcing the need for effective oversight and accountability mechanisms.
The methodology applied by Data & Justice is not limited to the situation in Maré nor to monitoring police violence. Its innovative approach can be replicated and adapted to different contexts to strengthen the protection of fundamental rights. In a world where marginalized communities face social exclusion and a lack of access to basic rights, initiatives like this represent a concrete way to promote greater transparency and access to justice.
Data & Justice demonstrates how technology can be an essential ally in the fight for transparency and social justice. Moreover, its replicable and adaptable nature allows it to expand beyond the monitoring of police violence. The approach can be applied in many other areas, such as reporting environmental violations, defending housing rights, and ensuring access to essential services.
Initiatives like Data & Justice reaffirm the importance of ethical technology use in addressing structural inequalities and strengthening the fight for fundamental rights. By employing artificial intelligence to make injustices more visible, this project paves the way for a new form of digital advocacy, where data becomes a tool for resistance and social transformation.