La incorporación de sistemas de decisión automatizada en las entidades públicas ha intensificado las tensiones entre la racionalidad algorítmica y la estructura dogmática del acto administrativo. En ese marco, el presente artículo tiene por objeto analizar las condiciones jurídicas bajo las cuales el uso de sistemas de decisión automatizada para asistir en la motivación del acto administrativo resulta compatible con los principios estructurales que rigen su validez. El análisis pone de manifiesto que dicha compatibilidad solo es posible cuando estos sistemas se subordinan a una arquitectura jurídica que garantice la determinación jurídica de los hechos, la motivación suficiente, la imputación a autoridad competente y el control efectivo, con supervisión humana, trazabilidad y responsabilidad institucional.
The integration of automated decision-making systems into public administration has intensified tensions between algorithmic rationality and the doctrinal structure of the administrative act. This article analyzes the legal conditions under which such systems may assist in the reasoning underlying administrative acts without undermining the structural principles governing their validity. The analysis demonstrates that such compatibility is possible only when automated systems are strictly subordinated to a legal framework that safeguards the legal determination of facts, the requirement to provide adequate reasons, the attribution of the act to a competent authority, and the availability of effective administrative and judicial review. Ensuring such subordination requires express statutory authorization, ex ante legal parameterization of system design and operation, meaningful human oversight, robust traceability throughout the decision-making process, and institutionally articulated responsibility. The article concludes that artificial intelligence cannot operate as an autonomous source of administrative decision-making, but only as a technical instrument integrated within the guarantees of the rule of law.