Decisiones automatizadas en la función pública en América Latina

Decisiones automatizadas en la función pública en América Latina (Automated Decisions in the Public Function in Latin America) is a Spanish-language comparative regional report co-authored by Patricio Velasco Fuentes and Jamila Venturini for Derechos Digitales, launched on 21 April 2021 as the comparative spine of Derechos Digitales' Inteligencia Artificial e Inclusión en América Latina programme line. The comparative report binds together six country case studies — two on Brazil (the National Employment System and pandemic-era Emergency Aid; Paula Cardoso), two on Chile (the Sistema Alerta Niñez child-rights predictive-risk algorithm and an urban-crime predictive system; Matías Valderrama), one on Colombia (the PretorIA project applying AI in constitutional-court jurisprudence; Víctor Saavedra), and one on Uruguay (pandemic-era health-information management; Dina Yael) — into a single Latin American mapping of automated decision-making in public services across the four jurisdictions named on the cover. The programme is funded by Canada's IDRC (International Development Research Centre) inside the Cyber Policy Centres (CPC) research network and the report and its country case studies are published under a CC BY 4.0 licence.

Methodology and central argument

The comparative report's central empirical claim is that automated decision-making technologies are becoming "increasingly common" in Latin American public administration — visible in social-protection, employment, criminal-justice, public-health, and constitutional-court administration — while regulatory frameworks, transparency obligations, and rights-protection capacity in the four jurisdictions all lag the pace of deployment. The methodology Velasco and Venturini install is a five-dimension human-rights audit applied to every case study: the national implementation context; the regulatory and institutional context; the data infrastructure underlying the system; the decision-making process the system actually executes; and the technological design itself. Each country chapter pairs a documented case (one of the six initial cases listed above) with a five-dimension analysis, and the comparative report extracts the cross-country patterns: a uniform absence of public registers of deployed systems, opaque procurement pipelines, weak or non-existent informed-consent infrastructure for the people whose administrative records feed training data, and an asymmetry between the speed of public-sector AI deployment and the speed of regional rights-protection scaffolding.

The Chile case study IA e inclusión: Chile — Sistema Alerta Niñez is the most-cited individual chapter in the corpus's own reading: it examines the Chilean Ministry of Social Development and Family's predictive-risk algorithm — designed to anticipate violations of children's rights — and raises concerns about the use of administrative data drawn from the Civil Registry, the Ministry of Education, and the Chile crece contigo social-protection programme, the risk of replicating administrative biases, and the absence of effective informed consent. The chapter is the load-bearing concrete example through which the comparative report's general framing (automated decision-making is being deployed faster than regulatory and rights-protection capacity) is anchored.

Programme position and authorship

Decisiones automatizadas en la función pública en América Latina is the comparative anchor of Derechos Digitales' Inteligencia Artificial e Inclusión en América Latina line — an ongoing line of work the organisation has run since 2019 to investigate how Latin American states are deploying AI-based and automated decision-making technologies. The line subsequently extended through four additional case studies on Argentina, Mexico, Paraguay, and a second Colombian system, applying the same five-dimension methodology. The April 2021 launch was structured around the comparative report and the six country case studies together — the comparative report is the single artefact that synthesises the case-by-case findings into a regional-pattern argument.

Authorship sits across two layers. The comparative report is co-authored by Patricio Velasco Fuentes (Derechos Digitales — analyst of regional patterns) and Jamila Venturini (Derechos Digitales and Al Sur). The country case studies are authored by Paula Cardoso (Brazil), Matías Valderrama (Chile), Víctor Saavedra (Colombia, on PretorIA), and Dina Yael (Uruguay), with academic collaboration on the Colombia case from Juan David Gutiérrez at the Universidad del Rosario. At time of publication Venturini was already serving as the programme lead on Derechos Digitales' AI-and-inclusion programme line and was several months from being elevated to the organisation's co-executive directorship in September 2021; the report sits inside her named public-output record on Latin American ADM mapping carried under her Voice entry.

Posture within the corpus

Within the corpus, Decisiones automatizadas en la función pública en América Latina is the first Latin American publication on the slate — closing the geographic gap that the prior publications had clustered around the United States (Unmasking AI, Bug Bounties for Algorithmic Harms, A Hazard to Human Rights, Losing Humanity, Comply To Fly?, On the Dangers of Stochastic Parrots), the United Kingdom (Participatory Data Stewardship, Not a drop to drink), and Continental Europe (Automating Society Report 2020). It is also the corpus's first Spanish-language publication, the first publication anchored on Global South public-sector AI deployment as its empirical object, and the first publication in the corpus to install the regional framing — most visible in the APC 20th-anniversary feature — that Latin America is itself "at the forefront of the assessment of impacts caused by artificial intelligence and data colonialism" rather than a region absorbing Global North critique.

The report's structural counterpart on the corpus's publications slate is Automating Society Report 2020, AlgorithmWatch's Continental European multi-country ADM mapping under Matthias Spielkamp's editorial direction. Both reports apply per-country case-by-case methodologies to public-sector ADM deployment, both pair empirical documentation with explicit policy framings addressed to regional rights-protection scaffolding (the EU AI Act on Spielkamp's side; the Inter-American Commission on Human Rights and UNESCO on Venturini's), and both anchor a recurring civil-society research programme rather than a one-off intervention. Together they form the corpus's regional-mapping evidence-base on its European and Latin American poles — the structural counterweight to the corpus's algorithmic-auditing pole (AJL's Gender Shades, Comply To Fly?, and Bug Bounties for Algorithmic Harms) and the long-form policy-plan pole (A Narrow Path).