From algorithm to justice: philosophical and ethical limits of justice automation
The article explores the issue of the correlation between justice and algorithmic automation of justice. The question of the nature of justice gains new dimensions in the context of the artificial intelligence use in judicial proceedings. Classical concepts of justice are compared with the capabilit...
Uloženo v:
| Vydáno v: | Аналітично-порівняльне правознавство Ročník 3; číslo 5; s. 449 - 453 |
|---|---|
| Hlavní autoři: | , |
| Médium: | Journal Article |
| Jazyk: | angličtina |
| Vydáno: |
05.11.2025
|
| ISSN: | 2788-6018, 2788-6018 |
| On-line přístup: | Získat plný text |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Shrnutí: | The article explores the issue of the correlation between justice and algorithmic automation of justice. The question of the nature of justice gains new dimensions in the context of the artificial intelligence use in judicial proceedings. Classical concepts of justice are compared with the capabilities of algorithms, which, despite their apparent impartiality, reflect the biases of their creators and training data. Special attention is paid to procedural justice and the problem of “fair procedure” requiring combining the result with the proper decision-making process. It is emphasized that algorithmic systems function in the paradigm of inductive generalization, which contradicts the principles of legal certainty and standards of proof. The hermeneutic tradition demonstrates the inability of algorithms to adequately take into account the context and value aspects. The paper also analyzes the concept of value pluralism of Isaiah Berlin, which indicates the impossibility of reducing moral values to mathematical criteria, which threatens reductionism in justice. The issue of the “accountability gap” is considered, since algorithmic systems are deprived of freedom and moral responsibility that are necessary conditions for the legitimacy of a judicial decision. Epistemological challenges are particularly emphasized, namely the opacity of algorithms contradicts the requirements of the validity and comprehensibility of judicial decisions. Attention is paid to the temporal aspect of justice, since algorithms trained on historical data risk reproducing outdated prejudices. Based on philosophical and legal analysis, the principle of algorithmic justice subsidiarity is substantiated: automation is possible in the field of procedural, analytical and standardized decisions, but is unacceptable for issues of guilt, punishment or complex value dilemmas. The principles of the ethical architecture of algorithmic justice are proposed, in particular moral agency, hermeneutic transparency, value completeness and temporal justice. It is concluded that the automation of justice can only act as an auxiliary tool, while the ultimate responsibility for the legal decision should belong to the human judge. |
|---|---|
| ISSN: | 2788-6018 2788-6018 |
| DOI: | 10.24144/2788-6018.2025.05.3.67 |