Authorship, detection, and judicial decision-making in the face of ai-assisted texts
DOI:
https://doi.org/10.64894/7kpqr194Keywords:
artificial intelligence, judicial authorship , judicial reasoning, human controlAbstract
This article examines the difficulties involved in identifying AI-assisted texts and their impact on judicial adjudication. Through a doctrinal and case-law-based approach, it analyzes, first, the conceptual and technical limits of detecting texts generated by language models and, second, the judicial decisions issued in the Chubut case. The study shows that the problem is not limited to AI detection or hallucinations, but rather concerns deeper issues related to judicial authorship, reasoning in judgments, traceability of legal reasoning, and effective human control. Its main finding is that the validity of a judicial decision depends less on the instrumental origin of the text than on its attributable authorship and jurisdictional control.
Downloads
References
BASANI, A. R. y CHEN, P.-Y. (2025). Diversity boosts AI-generated text detection. https:// arxiv.org/abs/2509.18880.
BORILE, C. y ABRATE, C. (2025). How to generalize the detection of AI-generated text: Confounding neurons. https://aclanthology.org/2025.findings-emnlp.1388.pdf. CENTRO DE ESTUDIOS DE JUSTICIA DE LAS AMÉRICAS [CEJA] (2025). Inteligencia artificial en los Poderes Judiciales: Reflexiones y lineamientos para las Américas. https://cejamericas.org/wp-content/uploads/2025/12/Inteligencia-Artificial-en los-Poderes-Judiciales-de-las-Americas_.pdf.
CONSEJO DE LA MAGISTRATURA DE LA CIUDAD AUTÓNOMA DE BUENOS AIRES [CM CABA] (2025). Nuevo Manual de Buenas Prácticas para el uso de IA e IAGen en el Po der Judicial de la CABA. Secretaría de Innovación.
GENG, M. y POIBEAU, T. (2025). On the detectability of LLM-generated text: What exactly is LLM-generated text? https://arxiv.org/abs/2510.20810.
MARTÍN, C., ALVARADO, G., RAMOS PEREYRA, L., SÁNCHEZ CAPARRÓS, M. y NUCCI TELLI, M. (2026). ¿Lo escribió un humano o la IA? Evaluación de autoría en textos ge nerados o asistidos por inteligencia artificial. IALAB - Facultad de Derecho, UBA.
Downloads
Published
Issue
Section
Categories
License
Copyright (c) 2026 Juan Manuel Garay (Autoría)

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.



