HHS · no agency ID · record R0600

Structuring Notice of Concern Data

Department of Health and Human Services · HHS/ACF

Overview

Development stage: DeployedSource: c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.
High-impact designation: Presumed high-impact, determined notSource: b) Presumed high-impact, but determined not high impact
Impact justification: The use of AI is narrowly focused on extracting key data points from unstructured narratives and validating data completeness. The outputs do not serve as a principal basis for decisions or actions with legal, material, binding, or significant effect on any of the 6 cases outlined in M-25-21, page 19.
Start date: 2024-12 (month precision)Source: Dec-24
Withheld from public reporting: NoSource: a) No

Mission

Topic area: Government Benefits Processing
Operational functions: Document processing and extraction, Risk scoring and triage Derived by keyword rules; see methodology

Problem the AI is intended to solve

How can the Office of Refugee Resettlement (ORR) clear its backlog of notices of concern (NOC) and minimize backlog in the future? Notice of Concern (NOC) forms contain critical information regarding safety of children who have left ORR's care. Some forms are received as scans, with the information not in machine-readable format. ORR receives hundreds of NOCs a day. Due to personnel shortage in the Prevention of Child Abuse and Neglect Team (PCAN) team responsible for reviewing and acting on NOCs, as of October 2024 there was a backlog of over 30,000 NOCs.

Expected benefits

More effective and efficient review of NOCs With AI-enabled structuring of data in NOCs received in scanned formats, ORR can reduce the large backlog that has accumulated.

System outputs

Structured data parsed from the subset of NOCs that are scans of documents AI is not used to triage NOCs, just to parse information from scanned documents. The parsed information is presented to the PCAN team alongside the original document for review and action.

Technology

AI classification: Computer visionSource: Computer Vision: AI that processes and interprets visual data (e.g., images and videos).
System name(s): ACF Horizon
Custom-developed code: Yes
Public source code: Not reported

Sourcing

How it was built: Contract and in-houseSource: c) Developed with both contracting and in-house resources
Vendor (as reported): Palantir
Vendors (standardized): Palantir

A vendor is the supplier named by the agency. It does not identify the underlying model or AI technology.

Data and privacy

Involves PII: Yes
Privacy Impact Assessment: www.hhs.gov ↗
Authorization to Operate: Yes
Demographic features: k) None of the above
Training and evaluation data: No training or fine-tuning; we are using secure commercially available LLMs.
Federal Data Catalog: Not reported

Governance

0 of 8 minimum-practice questions answered. A blank answer means the agency reported nothing; it does not mean the practice is absent.

Pre-deployment testingNo answer reported

AI impact assessmentNo answer reported

Independent reviewNo answer reported

Ongoing monitoringNo answer reported

Operator trainingNo answer reported

Fail-safeNo answer reported

Appeal processNo answer reported

User and public consultationNo answer reported

Potential impacts and how they were identified

Not reported