AI-Assisted Resident Outreach List Building for Social Services

What it is for

Eligibility & Public Benefits

PEU uses the system to identify and reach out to residents who may qualify for or benefit from public services, such as New Yorkers who recently turned 26 and may need help enrolling in health insurance.

NYC Algorithmic Tools Compliance Report — SmartVAN / TargetSmart

Run by: Department of Social Services (DSS)
Where: No fixed location
Kept: Retained Not specified; TargetSmart data is described as frequently updated and maintained as a standing dataset within SmartVAN
Shared with: Accountable organization
Your copy: You cannot see the data it holds about you. What you can do

What it is for

Eligibility & Public Benefits

The NYC Mayor's Public Engagement Unit uses SmartVAN with TargetSmart data to build lists of New York City residents for targeted outreach about social services such as health insurance enrollment. The system draws on a commercially purchased dataset of NYC residents including demographic, contact, and algorithmically derived likelihood scores. Residents may appear on outreach lists without having directly interacted with the city, based on commercially sourced data whose underlying algorithmic processes are proprietary to TargetSmart.

What it collects and what happens to it

Data taken in

Sensitive personal information

Identifiable data

A frequently updated commercial dataset from TargetSmart containing names, contact information, demographic data, and other personal information about New York City residents. Also includes data on existing clients or from partner agencies when available.

NYC Algorithmic Tools Compliance Report — SmartVAN / TargetSmart

About behaviour

Identifiable data

Survey responses and personal information collected from clients during prior outreach interactions, stored in SmartVAN for use in subsequent outreach projects.

NYC Algorithmic Tools Compliance Report — SmartVAN / TargetSmart

Processing

Classification & Prediction

TargetSmart uses proprietary algorithmic processes to match multiple input datasets to determine residency, contact information, and demographics, and to produce likelihood scores such as the probability that a household contains children under 18.

NYC Algorithmic Tools Compliance Report — SmartVAN / TargetSmart

What it does

Deciding (Analytical AI)

Human decides

TargetSmart's proprietary algorithms produce likelihood scores (e.g., likelihood that a household contains children under 18) and matching-derived demographic profiles. PEU staff use these scores and profiles to select outreach lists, but the final outreach decisions are made by PEU staff.

NYC Algorithmic Tools Compliance Report — SmartVAN / TargetSmart

Outputs

Sensitive personal information

Identifiable data

Lists of named, individually identifiable NYC residents with associated demographic and contact information, generated for PEU outreach campaigns. These outputs are the result of TargetSmart's proprietary algorithmic matching and scoring processes.

NYC Algorithmic Tools Compliance Report — SmartVAN / TargetSmart

A recommendation or prediction

Identifiable data

Algorithmically derived likelihood scores (e.g., likelihood that a household contains children under 18) that guide which residents are prioritized for outreach. These scores are produced by TargetSmart's proprietary processes and are advisory inputs to PEU list selection.

NYC Algorithmic Tools Compliance Report — SmartVAN / TargetSmart

Run by

Department of Social Services (DSS)

The Mayor's Public Engagement Unit (PEU) within the NYC Department of Social Services deploys SmartVAN to manage outreach across a range of projects, using TargetSmart data to build lists of residents and collect survey responses.

NYC Algorithmic Tools Compliance Report — SmartVAN / TargetSmart

Built by

EveryAction and TargetSmart

EveryAction is the software provider of the SmartVAN platform. TargetSmart is the data provider that applies algorithmic techniques to derive residency, contact, and demographic data on NYC residents. EveryAction provides access to TargetSmart data through their platform.

NYC Algorithmic Tools Compliance Report — SmartVAN / TargetSmart

Kept for

Retained Not specified; TargetSmart data is described as frequently updated and maintained as a standing dataset within SmartVAN

The TargetSmart commercial dataset is a frequently updated standing dataset within SmartVAN. Survey responses and client data collected during outreach are also stored within the platform. Specific retention periods are not disclosed in the register entry.
Duration: Not specified; TargetSmart data is described as frequently updated and maintained as a standing dataset within SmartVAN

NYC Algorithmic Tools Compliance Report — SmartVAN / TargetSmart

Shared with

Available to the accountable organization

PEU staff access the output lists and likelihood scores within the SmartVAN platform. The underlying input data and algorithmic processes used to generate the data remain proprietary to TargetSmart and are not directly accessible to PEU.

NYC Algorithmic Tools Compliance Report — SmartVAN / TargetSmart

Not available to me

Individual residents cannot access, review, or correct the TargetSmart data held about them within SmartVAN, nor can they see the likelihood scores assigned to their household. No individual access mechanism is described in the register entry.

NYC Algorithmic Tools Compliance Report — SmartVAN / TargetSmart

Stored

Stored on 3rd Party Cloud

Data is stored and managed on the EveryAction SmartVAN platform infrastructure, which is a third-party cloud service. The TargetSmart commercial dataset is hosted within this platform. Specific storage jurisdiction is not described in the register entry.
Duration: Not specified; data is maintained as a standing commercial dataset

NYC Algorithmic Tools Compliance Report — SmartVAN / TargetSmart

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 Be Informed of AI UseResidents who appear on PEU outreach lists have a right to be informed that an AI system contributed to their selection. The register does not describe a proactive notice mechanism; residents contacted by PEU may not know they were identified through TargetSmart's algorithmic processes.
Right to Algorithmic TransparencyResidents have a right to understand how the algorithmic system that identified them for outreach works. The register acknowledges that both input data and algorithmic processes are proprietary to TargetSmart, significantly limiting the transparency the city can provide. Contact the NYC Department of Social Services for further information.

Risks and safeguards

Civil liberties harmResidents appear on outreach lists based on commercially sourced, proprietary algorithmic processes they cannot inspect or contest, raising concerns about chilling effects, surveillance, and consent. The register notes training data and input data are part of vendor proprietary processes, limiting transparency. Mitigation appears limited to restricting use to outreach (not enforcement) contexts; no opt-out or notice mechanism is described in the register entry.
Reputational harmTargetSmart's matching algorithms could misattribute residency, demographics, or household composition to individuals, resulting in incorrect targeting or stigmatizing categorizations. The proprietary nature of the input data and processes means errors are difficult to detect or correct. No described mitigation addresses individual-level accuracy review or correction procedures.