Protection of human rights in the context of implementing algorithmic technologies in pre-trial investigation

Abstract

This article provides a comprehensive and multi-layered analysis of the theoretical, legal, doctrinal, and ethical challenges determined by the systemic integration of algorithmic solutions and artificial intelligence (AI) technologies into the architecture of the modern criminal justice system. The author posits that the digital transformation of pre-trial investigation and judicial proceedings has catalyzed a fundamental ontological conflict between the paradigm of technological determinism – which prioritizes the maximization of procedural efficiency through the automated processing of vast datasets – and the classical anthropocentric legal paradigm, which asserts the absolute priority of fundamental human rights, individual agency, and personhood. This tension necessitates a radical re-evaluation of the conceptual foundations of criminal procedure in the era of the Fourth Industrial Revolution.

The research focuses on the epistemological risks inherent in the use of “black-box” algorithms within law enforcement, where the opacity of mathematical models utilized in predictive policing and recidivism risk assessment systems threatens the core principle of legal certainty. The author substantiates the necessity of a conceptual transition toward a new doctrine of “digital humanism.” This doctrine is envisioned as a sophisticated synthesis of institutional legal safeguards and ethical engineering filters, implemented through the innovative framework of “Ethics by Design.” It is argued that ethical norms must be embedded directly into technical protocols and the underlying source code of software utilized in criminal proceedings, thereby transforming algorithmic architecture into an accountable and transparent tool of justice rather than a self-governing entity.

A pivotal component of this study is the critical examination of the mathematical and socio-legal limitations of equity in automation. Specifically, the article explores the “Kleinberg-Chouldechova impossibility theorem of fairness,” which demonstrates that different definitions of algorithmic equity – such as calibration, predictive parity, and error rate balance – are mathematically incompatible under certain conditions. The author argues that this theorem serves as a crucial warning against over-reliance on purely technical solutions to social bias. It underscores the fact that achieving “fairness” in a criminal justice context is not merely a computational task but a profound political and legal choice that requires human judgment to navigate the inherent trade-offs between competing metrics of equality.

Furthermore, the article addresses the transformative evolution of conventional human rights in the digital age. The author provides a detailed analysis of the genesis and substantive content of the “right to an explanation” regarding algorithmic logic, which is emerging as an indispensable condition for ensuring the right to a fair trial. It is demonstrated that without the ability of the defense to verify, challenge, and rebut algorithmic outputs, the adversarial character of the judicial process is fundamentally compromised. Additionally, the study investigates the persistent problem of algorithmic bias arising from the use of unrepresentative or historically skewed training data. The author asserts that protection against “automated stigmatization” must attain the status of a specialized procedural guarantee to prevent the perpetuation of systemic social prejudices through seemingly objective digital instruments.

In its concluding remarks, the article emphasizes that the legitimacy of innovative investigation and adjudication depends not on the technical perfection of algorithms, but on the legal system’s capacity to ensure meaningful human oversight – the “human-in-the-loop” principle. The author concludes that the future of criminal procedural doctrine must focus on enhancing the transparency, explainability, and auditability of algorithmic systems. This study holds theoretical significance for the development of legal doctrine in the digital epoch and offers practical insights for the establishment of rigorous validation standards for forensic and investigative software, ensuring that technological progress serves the interests of truth and human dignity.