Supervisión humana
Revisión y corrección humana de decisiones algorítmicas.
How the laws define this
Public sector governance frameworks consistently define human oversight as mandatory human involvement to ensure human judgment and accountability remain central when using artificial intelligence in public administrative decisions. The primary point of divergence concerns when and how this intervention must occur. The Buenos Aires Province Generative AI Guidelines require continuous human intervention and oversight specifically before AI-generated outputs are used in administrative decisions. In contrast, Resolución 4/2025 - Directrices de uso de Inteligencia Artificial Generativa en la Administración Pública de la Provincia de Buenos Aires mandates human control and oversight both before and after the use of AI systems. Meanwhile, Proyecto de Ley No. 417 de 2025, which regulates AI in managing petitions, complaints, claims, suggestions, and reports in Colombian public entities, focuses on a human-in-the-loop requirement that automated decisions be reviewable, validable, and correctable by a competent public official, explicitly adding duties to provide review channels, recourse, and reparation when processing affects rights or benefits.
Synthesised from the 3 statutory definitions below. Each one is quoted in full, with its source.
How jurisdictions differ
What is distinctive in each definition, binding law in force first. Showing 2 of the 3 instruments that define this term; all are listed under Definitions below.
| Jurisdiction | Instrument | Status | What is distinctive |
|---|---|---|---|
| Argentina | Resolution 4/2025 - Guidelines for the Use of Generative Artificial Intelligence in the Public Administration of the Province of Buenos Aires | In Force | Adds timing by requiring human control both before and after AI use, and emphasizes preserving human judgment and accountability in decision-making. |
| Colombia | By which the use of Artificial Intelligence in the management of Petitions, Complaints, Claims, Suggestions and Reports (PQRSD) in public entities of the Colombian State is regulated and implemented, and other provisions are issued (Bill No. 417 of 2025) | Withdrawn | Adds requirements for reviewable AI decisions by competent public officials, including human-in-the-loop processes and mechanisms for appeals and remedies when rights are affected. |
What it triggers
- In the public administration of the Province of Buenos Aires, human control and oversight is required both before and after AI use to ensure human judgment and accountability in decision-making.source
- Under Bill No. 417 of 2025 in Colombia, AI systems processing PQRSD would have required human review, validation, correction, and appeal mechanisms by competent public officials when rights or benefits are affected.source
Definitions (3)
Requisito de que las decisiones generadas por sistemas de IA sean revisables, validables y corregibles por un funcionario público competente (human-in-the-loop), incluyendo la obligación de habilitar canales de revisión humana y mecanismos de recursos y reparación cuando el procesamiento automático afecte derechos o prestaciones.
The requirement for continuous human intervention and oversight in the review of AI-generated outputs before they are used in administrative decisions.
The requirement for human control and oversight both before and after the use of AI systems, ensuring that human judgment and accountability remain central to decision-making processes.
Not to be confused with
Control humano significativo focuses on meaningful oversight of AI outcomes, while Supervisión humana here specifies human control before and after use or continuous review of outputs.
Human supervision is a general requirement for oversight of AI decision-support systems, whereas Supervisión humana adds details like timing, intervention, or appeal mechanisms.
Significant and competent human supervision emphasizes understanding and overriding AI outputs, differing from Supervisión humana's focus on review processes, timing, or remedies.
Questions and answers
What does human supervision require under Buenos Aires Province rules?
Resolution 4/2025 requires human control and oversight both before and after AI use to keep human judgment central. The guidelines recommend continuous human review of AI outputs before administrative use.
Is human supervision mandatory for AI in Colombian public entities?
Bill No. 417 of 2025, which was withdrawn, would have required AI decisions in PQRSD processing to be reviewable and correctable by competent officials with appeal mechanisms.
How does Supervisión humana differ across these definitions?
Definition 1 covers pre- and post-use oversight emphasizing accountability; definition 2 focuses on pre-use review of outputs; definition 3 mandates reviewability and remedies by officials.
Related Terms
Control humano significativo
Meaningful human oversight over AI system outcomes....
Human supervision
Requirement for human oversight of AI decision-support systems....
Significant and Competent Human Supervision
Human oversight enabling understanding and override of AI outputs....
Human agency and oversight
Requirement to preserve meaningful human decision‑making and control....
Human-in-the-loop (Human Oversight)
Design ensuring human oversight and intervention....