Algorithmic governance in the Judiciary and its application to environmental proceedings: empirical evidence from the CNJ and proposition of the Operational Protocol for Algorithmic Environmental Evidence

Abstract

This study examines the conditions for the ethical and procedurally secure use of Generative Artificial Intelligence within the Brazilian Judiciary, with a specific focus on environmental criminal prosecution. It investigates the extent to which the normative framework established by the National Council of Justice provides sufficient safeguards for the use of algorithmic environmental evidence and how such requirements may be operationalized in a verifiable manner. The analysis is grounded in the premise that, although the current regulatory framework is based on principles of algorithmic governance, it still lacks an operational systematization capable of ensuring the traceability of inferences and the effective exercise of technical adversarial review. The research adopts a qualitative and exploratory approach, based on documentary analysis of normative acts and institutional reports issued by the National Council of Justice, within the context of contemporary national judicial policy. The findings indicate that the expansion of algorithmic systems in the Judiciary is accompanied by risks related to opacity, bias, and weaknesses in the digital chain of custody, with significant impacts on evidentiary integrity in environmental cases. As a contribution, the study proposes an Operational Protocol for Algorithmic Environmental Evidence, aimed at bridging the gap between abstract regulation and the practical application of generative artificial intelligence in environmental criminal litigation.