Answers are grounded in this paper and cite their sources.
This article examines how the adoption of artificial intelligence (AI) in the overt policing carried out by the Pará Military Police, especially automatic license plate reading and facial recognition, relates to Brazilian law and to evidence on algorithmic bias. The qualitative, exploratory and descriptive study relied on bibliographic and documentary procedures and examined releases from the State Department of Public Security, the force's statutes, federal legislation and precedents of the Superior Court of Justice. The results indicate that the declared network grew from about 200 cameras without AI in 2018 to over a thousand cameras with AI in 2026, with no disclosed error rates, data protection impact assessment or protocol for stops resulting from alerts, a gap made more serious by a state whose population is mostly of mixed race. The study concludes that an algorithmic alert must be treated as a lead subject to recorded human confirmation, and proposes seven minimum governance parameters for a state normative act and for the force's activity.
No takes yet. Share an insight, caveat, or question.
Souza et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: