AI-Assisted Child Welfare Caseload Forecasting for NYC

What it is for

Planning & Decision-making

The model forecasts weekly child welfare investigation caseloads by borough for the next 12 months, enabling ACS to plan how many Child Protective Service social workers to assign to each borough field office as workers complete their training.

NYC Algorithmic Tools Compliance Report — Caseloads Projection

Employment & Work

Staffing decisions — specifically the assignment of newly trained Child Protective Service social workers to specific borough field offices — are made based on the projected caseloads output by this model, combined with historical attrition rates and current team size.

NYC Algorithmic Tools Compliance Report — Caseloads Projection

Run by: Administration for Children's Services (ACS)
Where: No fixed location
Kept: Not stated by the NYC Office of Technology and Innovation.
Shared with: Accountable organization
Your copy: You cannot see the data it holds about you. What you can do

What it is for

Planning & Decision-making

Employment & Work

This system uses a time series forecasting model to predict the number of open child welfare investigations in each New York City borough over the next 12 months. The forecasts help the Administration for Children's Services assign the right number of social workers to each borough based on expected workload. No personal information about individual families or children is used — predictions are based solely on anonymised counts of investigations. Social workers are assigned based on these projections, historical attrition, and current team sizes.

What it collects and what happens to it

Data taken in

Operational data

Anonymized data

Administrative data on open child welfare investigations and Family Service Units' involvement, extracted from the CONNECTIONS child welfare computer system. A snapshot of all open investigations is taken each Monday and used as the average weekly caseload count. No identifying information about individual families or children is collected or used.

NYC Algorithmic Tools Compliance Report — Caseloads Projection

Processing

Classification & Prediction

A time series forecasting model that accounts for seasonality, trend, and variation in historic caseload data. Trained on ACS historic caseloads from July 2021 to December 2024 and tested on January–November 2025 data. Produces weekly forecast estimates at the borough level for the next 12 months from each run date.

NYC Algorithmic Tools Compliance Report — Caseloads Projection

What it does

Deciding (Analytical AI)

Human decides

The model produces borough-level caseload forecasts (scores and rankings of predicted workload). Staffing decisions are then made by human managers using those forecasts alongside attrition data and current team sizes — the model is explicitly informative, not decision-making.

NYC Algorithmic Tools Compliance Report — Caseloads Projection

Outputs

Operational data

Anonymized data

Forecast of the number of open child welfare investigations involving children, broken down by NYC borough, for each week over the next 12 months. These aggregate counts contain no personal information and are used internally by ACS management for workforce planning.

NYC Algorithmic Tools Compliance Report — Caseloads Projection

Run by

Administration for Children's Services (ACS)

The Administration for Children's Services (ACS) deploys and operates this caseload forecasting model through its Division of Policy, Planning, and Analysis. ACS is responsible for the model's training, testing, and quarterly use in staffing decisions.

NYC Algorithmic Tools Compliance Report — Caseloads Projection

Built by

Not stated by the NYC Office of Technology and Innovation.

Kept for

Not stated by the NYC Office of Technology and Innovation.

Shared with

Available to the accountable organization

Forecast outputs are used internally by ACS management (Division of Policy, Planning, and Analysis) to guide staffing allocation decisions. The register does not indicate that outputs are shared externally.

NYC Algorithmic Tools Compliance Report — Caseloads Projection

Not available to me

Members of the public, including families and children involved in investigations, cannot access the model's forecast outputs. The system produces aggregate workforce-planning data for internal ACS use only.

NYC Algorithmic Tools Compliance Report — Caseloads Projection

Stored

Not stated by the NYC Office of Technology and Innovation.

How to read the colours

Can it identify you?

Anonymized data: Data about people with the link to who is broken. Stripped of identifiers, blurred, aggregated, or noised so this system can’t reasonably tie a record back to an individual.
Pseudonymous data: Each person’s data is tied to a token (hash, ID, template) that lets this system recognise the same person across events, but the token itself doesn’t reveal a name. Reidentification is possible with extra information.
Identifiable data: The data either contains a direct identifier (name, address, account name, recognisable face or voice, plate number) or carries a token this system uses to look up legal identity during processing.

Who completes the loop?

Human decides: This mode suggests; a person decides what to do next. The AI is always advisory — a human is in the loop on every decision. Example: a triage tool ranks cases for a clinician who chooses which to see first.
Human executes: This mode decides; a person carries out the result. Example: an optimizer plans the day’s trash-collection routes, and drivers run them.
Autonomous: This mode decides and acts on its own. No person reviews each decision or carries out the resulting action.

Definitions from the DTPR standard. Amber is about your data, violet about who decides. The fuller the shape and the deeper the colour, the more identifying the data or the less a person is involved.

What you can do

Ask about this system

Questions go to the NYC Office of Technology and Innovation, not the vendor.

Your rights

Right to Algorithmic TransparencyMembers of the public can learn about this system through the NYC Algorithmic Tools Compliance Report published at data.cityofnewyork.us. The report describes how the model works, what data it uses, and how its outputs are used in staffing decisions. No individual-level decisions are made by this system, so there is no personal decision to contest or explain.

Risks and safeguards

Financial & business harmInaccurate caseload forecasts could lead to understaffing in high-need boroughs, leaving vulnerable children without timely investigation, or to overstaffing causing inefficient use of public resources.Safeguard: ACS retrained the model on data through December 2024 and tested it against January–November 2025 actuals before operational use. The model output is one input among several — managers also use historical attrition rates and current team sizes, preserving human judgment in the allocation decision.