evidoria

← Back to browse

Good practice Imported

Repositorio de Algoritmos Públicos — GobLab UAI's Five-Year Public Audit of Chile's State Algorithms

Chile · Santiago · See the Chile profile · See the Santiago profile

Evidence: Observational / pre–post Top 14% 80/100 · Ask Evidence Copilot about this practice

Since 2021, Universidad Adolfo Ibáñez's GobLab has independently catalogued automated decision systems used by the Chilean state. Its 2026 report documents 162 systems — up from 75 in 2023 — and underpins Chile's first Latin American algorithmic-transparency directive.

162
Automated decision systems catalogued (2026 report)
75
Systems catalogued in the first report (2023 report)
43
New systems added in the latest report (2026 report)
800+
Public bodies surveyed for the 2023 transparency instruction
13
Institutions publishing their own disclosures using GobLab's toolkit (by 2025)
Repositorio de Algoritmos Públicos — GobLab UAI's Five-Year Public Audit of Chile's State Algorithms

Details

Maturity
Established
Promoter
GobLab, Escuela de Gobierno, Universidad Adolfo Ibáñez, with Consejo para la Transparencia
Period
2021–2026
Keywords
public administration, algorithmic transparency, AI governance, digital government

Context

GobLab, the government-innovation laboratory of the School of Government at Universidad Adolfo Ibáñez, has independently identified and catalogued automated and semi-automated decision systems used by the Chilean state since 2021, publishing an annual report on its Repositorio de Algoritmos Públicos. Systems are identified through a documented methodology and reviewed by a multidisciplinary editorial committee, and the project works in partnership with Chile's official Consejo para la Transparencia (CPLT).

Objectives

Build public, independent visibility into where and how the Chilean state uses automated decision systems, and use that evidence base to support algorithmic-transparency policy.

Activities

The work underpins Chile's 2023 General Instruction on Algorithmic Transparency, developed with CPLT and the first binding instrument of its kind in Latin America, which surveyed more than 800 public bodies. In 2025, 13 institutions - including the National Employment and Training Service (SENCE) and the Social Insurance Institute (IPS) - applied GobLab's ethical-algorithms toolkit and subsequently published their own algorithm disclosures.

Results

Documented coverage has grown every year: 75 systems in the 2023 report, 101 in 2024 (26 new), 117 in 2025 (16 new), and 162 in the 2026 report (43 new) - five consecutive editions.

Conclusions

The repository remains an independent academic monitoring effort rather than a legally binding registry: of the more than 800 public bodies surveyed for the 2023 instruction, only a small subset (13 institutions by 2025) have gone on to publish their own disclosures using GobLab's tools, so most of the Chilean state's algorithmic use is still documented by GobLab's own research rather than by agencies' self-reporting.

Implementation

Indicative cost
Low (< €50k)
Time to results
Long (> 3 years)
Staffing & skills
GobLab research team at Universidad Adolfo Ibáñez's School of Government, Multidisciplinary editorial committee reviewing identified systems, Partnership with Chile's Consejo para la Transparencia (CPLT)

Conditions for success

  • A documented, repeatable methodology for identifying and reviewing systems each year
  • Formal partnership with the official transparency oversight body (CPLT)
  • Practical toolkits (ethical-algorithms toolkit) that public institutions can apply to publish their own disclosures

Commonly funded by

CERV — Citizens, Equality, Rights and Values Horizon Europe National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Replication kit

Reusable artefacts from this practice — as published by their sources.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful