PAN Lab example

Sistema Alerta Niñez (Chile)

Scored before anyone knocks: a child-risk targeting tool

A model scores every child in the system for future risk and hands OLN teams a ranked list of whom to reach out to first — built from data families gave to receive benefits. Modeled on Chile's Sistema Alerta Niñez. Officially the score is 'one more input.' Watch the input boundary, where consented data becomes a risk ranking nobody was told about, and the review layer that was built but never turned on.

Stylized model of a documented deploymentChild welfare & family services

Open this example in PAN Lab v0.1 to apply pressures and levers and watch what the system does.

What this models

This example runs on the SAN-class predictive child-risk targeting tool network: 6 components and 10 pathways between them. Every context in the Lab is a stylized model, never a reconstruction of any actual deployment, and each assumption behind it carries a provenance label.

Evidence base: 5 assumed · 1 published baseline. In the Lab, the shaded evidence band behind each headline readout draws its width from the least-established class below.

assumed
This models the predictive-targeting pattern documented in the Sistema Alerta Niñez case file — not a reconstruction of the actual instrument.
baseline
The cross-agency-data node carries the model's real feature inflow — the 280 administrative variables the documentation describes — and is marked privacy-sensitive because that data was supplied by families for benefit eligibility, not for risk scoring, without informed consent to the ranking or a way to opt out.
assumed
The field-feedback pathway from the case platform into the model starts empty: the documentation records that OLN's territorial knowledge is not fed back into the predictive model, so the corrective loop is present on the map but carries nothing — a lever can open it, though doing so naively risks the documented performative feedback loop.
assumed
The independent bias audit is drawn as an oversight node whose check pathway starts closed: an audit funded and conducted but whose criteria and results were never made public delivers no corrective relief on the diagram, matching the documented account.
assumed
The priority-ranking worklist is drawn as a mediating artifact on the model → OLN pathway, reflecting the case file's account of a system whose output was a prioritized list. It carries no flow of its own and does not affect the dynamics; the score is officially subordinate to OLN judgment, so the adoption baseline is moderate rather than saturating.
assumed
Documented concern about a socioeconomic gradient in who scores highest is the developers' own qualitative acknowledgment, not a measured disparity published by an independent audit. This Lab models institutional propagation, not demographics, and estimates no differential harm to served children.

What this example does not show

The documented concern is a socioeconomic gradient in who scores highest — a pattern the developers acknowledged, not a measured disparity published by an independent audit. The Lab models institutional propagation, not demographics, and estimates no differential harm to served children; that concern is documented in the case file and measured outside any diagram like this one.

Sources and evidence

What this example rests on, claim by claim. Every entry resolves to the same ledger the Evidence Registry publishes.

Sistema Alerta Niñez drew on 280 administrative variables that families had supplied to access social benefits, without informed consent to the risk ranking or a way to opt out; the model's developers acknowledged it was less able to identify higher-income children at risk, because lower-income families have more contact with the state.
empirical
- Investigative Derechos Digitales (Matias Valderrama), IA e inclusion: Chile 'Sistema Alerta Ninez' y la prediccion del riesgo de vulneracion de derechos de la infancia (2021) https://www.derechosdigitales.org/wp-content/uploads/CPC_informe_Chile.pdf
- Investigative Derechos Digitales (Matias Valderrama), AI and Inclusion: Chile 'The Child Alert System' (2022) https://www.derechosdigitales.org/wp-content/uploads/02_Informe-Chile-EN_180222.pdf
- Academic Center for Human Rights and Global Justice, NYU School of Law (Victoria Adelmant), Risk Scoring Children in Chile (2022) https://chrgj.org/2022-04-20-risk-scoring-children-in-chile/

Where this connects

Institutional pressures in this domain

Workload surge — Demand outruns staffing; per-case attention shrinks and review becomes triage.
Deadline pressure — Statutory or managerial timeliness rules reward fast approval of machine output over slow disagreement.
Staff turnover — Experienced skepticism leaves; new staff calibrate their trust on the tool itself.
Data & policy drift — The world, the intake process, and the rules change under a system trained on how things used to be — two mechanisms with different remedies: the statistical properties of what the system processes move (concept drift), or the mixture of inputs arriving in deployment differs from the mixture it was trained on (covariate shift).
Compliance over substance — Paper controls (sign-offs, checklists) satisfy audits while the behavior they describe erodes.

All of them in context on the Child welfare & family services domain page.

Levers available here and the patterns behind them

Escalate checks — State-feedback vigilance
Vet connections — Connection authorization
Store less data — Data minimization
Gate vendor updates — Vendor quality gate
Assign a challenger — Structured dissent
Review on schedule — Oversight cadence & retrospectives
Require sign-off — Conformity assessment gate
Keep skills sharp — Deskilling-arrest mandate
Review the riskiest first — Risk-tiered oversight
Mark AI-written records — Provenance labeling
Upgrade model — Improve the model