USA-006

Data Mining Solution (DMS) for CalWORKs Stage 1 Child Care Program Fraud Detection

Download PDF
United States North America High income Full Production Deployment Confirmed

Los Angeles County Department of Public Social Services (DPSS); Los Angeles County Chief Executive Office – Service Integration Branch (SIB)

At a Glance

What it does Anomaly and change detection — Compliance and integrity
Who runs it Los Angeles County Department of Public Social Services (DPSS); Los Angeles County Chief Executive Office – Service Integration Branch (SIB)
Programme CalWORKs Stage 1 Child Care Program
Confidence Confirmed
Deployment Status Full Production Deployment
Key Risks Governance and institutional oversight risks
Key Outcomes Public reporting on the pilot and early operational period states that the system identified collusive fraud rings with an 85 percent hit rate and generated an estimated $6.
Source Quality 5 sources — Report (government / official), News article / media

The Data Mining Solution (DMS) is a fraud-detection and case-prioritisation system used by the Los Angeles County Department of Public Social Services (DPSS) in the CalWORKs Stage 1 Child Care Program. CalWORKs (California Work Opportunity and Responsibility to Kids) provides child care subsidies to welfare recipients so they can attend work or training. The DMS was implemented with SAS technology under Los Angeles County Agreement 77217 and is described in county, trade-press, and vendor-adjacent materials as a data-mining platform for surfacing suspicious cases for investigator review. The system was developed by SAS Institute and customised for DPSS, with the initial development and implementation costing approximately $2.4 million, not counting additional operational and maintenance fees. The software and hardware components, along with cloud-based hosting, are handled by SAS.

Publicly available reporting indicates that the system combines anomaly screening, rules-based checks, and social-network-style link analysis across programme case records. Using complex algorithms, the system generates risk scores derived from behavioural anomalies in child-care service usage, similar in concept to a credit score, which alert investigators to suspicious activities. The social network analysis component maps connections between similar names, phone numbers, bank accounts, and other data links between individuals who may be involved in a large fraud operation, allowing investigators to visualise these relationships graphically. The system integrates approximately 150 different data sources to find linkages between them. Documented fraud cases include false employment claims of nonexistent employees, heads of fraud rings colluding with recipients who falsely declare their children are attending nonexistent child care centres, and criminals declaring false or shorter work schedules than the time actually claimed.

The project originated from a 2007 Board of Supervisors vote to study data-mining technologies following published reports detailing widespread fraud in federal, state, and local government health and welfare programmes. A pilot phase conducted in 2008-2009 identified collusive fraud rings with a reported 85 percent hit rate and generated an estimated $6.8 million in cost avoidance, which supported a larger implementation contract with SAS. The Board of Supervisors approved the full DMS contract in December 2009. The system went live officially around 2011, and in its first ten months of operation generated 197 additional child-care fraud referrals and 67 additional non-child-care referrals.

Operationally, DMS sits upstream of human investigation rather than replacing it. The system produces alerts and link-analysis outputs that are reviewed by designated county staff, who decide whether to refer a matter to fraud investigators for formal inquiry. County investigators use the probability scores to help prioritise their caseload. Investigators from DPSS were involved in design sessions for the system from the outset, providing developers with precise detail on how to construct a user-friendly interface. The system also includes a mapping utility that helps investigators get an early look at unfamiliar areas they need to travel to for a case. Investigators, not the system, determine whether fraud occurred and whether any benefit, recovery, or enforcement action should follow.

The public record indicates that by 2012 DMS had delivered more than 200 total fraud referrals, with approximately 10 to 15 cases per month being referred and more than 40 percent of those referrals proving positive for fraud. Social network analysis uncovered two conspiracy rings comprising 16 cases significantly earlier than traditional methods would have surfaced them. The DMS concept was also extended beyond child care into In-Home Supportive Services, supported by a $10 million state funding allocation approved in October 2009 for IHSS fraud prevention. However, the most clearly documented and attributable operational record in the retained source pack remains the CalWORKs Stage 1 child care deployment.

Classifications follow the DCI AI Hub Taxonomy. Hover over field labels for definitions.

Social Protection Functions

Implementation/delivery chain
Accountability mechanisms primaryProvision of payments/services
SP Pillar (Primary) The social protection branch: social assistance, social insurance, or labour market programmes. Social assistance
Programme Name CalWORKs Stage 1 Child Care Program
Programme Type The type of social protection programme, classified under social assistance, social insurance, or labour market programmes. View in glossary Fee waivers and targeted subsidies
System Level Where in the social protection system the AI is applied: policy level, programme design, or implementation/delivery chain. View in glossary Implementation/delivery chain
Programme Description CalWORKs (California Work Opportunity and Responsibility to Kids) Stage 1 Child Care Program provides subsidised child care to welfare recipients in Los Angeles County to enable participation in work or training activities. This case focuses on the DMS integrity-screening deployment documented around the child care programme, rather than trying to comprehensively document later anti-fraud extensions in other county programmes.
Implementation Type How the AI output is produced: Classical ML, Deep learning, Foundation model, or Hybrid. Affects validation, compute requirements, and governance profile. View in glossary Classical ML
Lifecycle Stage Current stage in the AI lifecycle, from problem identification through to monitoring, maintenance and decommissioning. View in glossary Monitoring, Maintenance and Decommissioning
Model Provenance Origin of the AI model: developed in-house, adapted from open-source, commercial/proprietary, or accessed via third-party API. View in glossary Commercial/proprietary
Compute Environment Where the AI system runs: on-premise, government cloud, commercial cloud, or edge/device. View in glossary Commercial cloud
Compute Provider The specific cloud or infrastructure provider hosting the AI system. SAS Institute Inc. (Cary, North Carolina data centres)
Sovereignty Quadrant Classification of data and compute sovereignty: I (Sovereign), II (Federated/Hybrid), III (Cloud with safeguards), or IV (Shared Innovation Zone). View in glossary III — Compute-Intensive Cloud with safeguards
Data Residency Where the data used by the AI system is stored: domestic, regional, or international. View in glossary Domestic
Cross-Border Transfer Whether data crosses national borders, and if so, whether documented safeguards are in place. View in glossary None
Decision Criticality The rights impact of the decision the AI supports. High criticality requires HITL oversight; moderate requires HOTL; low may operate HOOTL. View in glossary High
Human Oversight Type Level of human involvement: Human-in-the-Loop (active review), Human-on-the-Loop (monitoring), or Human-out-of-the-Loop (periodic audit). View in glossary HOTL
Development Process Whether the AI system was developed fully in-house, through a mix of in-house and third-party, or fully by an external provider. View in glossary Mix of in-house and third-party
Highest Risk Category The most significant structural risk source identified: data, model, operational, governance, or market/sovereignty risks. View in glossary Governance and institutional oversight risks
Risk Assessment Status Whether a formal risk assessment, informal assessment, or independent audit has been conducted for this system. Not assessed

Risk Dimensions

Governance and institutional oversight risks
Market, sovereignty and industry structure risks

Impact Dimensions

Autonomy, human dignity and due process
  • Human oversight protocol
CategorySensitivityCross-System LinkageAvailabilityKey Constraints
Beneficiary registries and MISSensitiveLinks data across multiple systemsCurrently available and usedCase and fraud-related data from the CalWORKs Stage 1 Child Care and IHSS programmes, including fraud referrals and programme case records; approximately 150 different data sources integrated
Financial and payments data: programme operationsSensitiveLinks data across multiple systemsCurrently available and usedTransaction data used for social network analysis to map connections between participants and providers; includes payment records and billing data

County of Los Angeles DPSS / Office of the CIO (2012) 'Approve Amendment Number Two to Agreement 77217 with SAS Institute Inc. for Data Mining Solution (DMS)', CIO Analysis Report CIO 12-07. Los Angeles: County of Los Angeles. Available at: https://file.lacounty.gov/SDSInter/bos/supdocs/68310.pdf (Accessed: 23 March 2026 – connection error).

View source Report (government / official)

GovTech (2012) 'Los Angeles County Uses Analytics to Stop Child-Care Fraud', Government Technology, 16 May. Available at: https://govtech.com/health/Los-Angeles-County-Uses-Analytics-to-Stop-Child-Care-Fraud.html (Accessed: 23 March 2026).

View source News article / media

GovTech / SAS (2015) 'LA County Department of Public Social Services uses analytics to fight child care benefits fraud', Industry Insider California, 3 December. Available at: https://insider.govtech.com/california/sponsored/la-county-department-of-public-social-services-uses-analytics-to-fight-child-care-benefits-fraud.html (Accessed: 23 March 2026).

View source News article / media

SAS Institute (2012) 'Social Network Analysis and the government fraudster', SAS State and Local Government Blog, 20 September. Available at: https://blogs.sas.com/content/statelocalgov/2012/09/20/social-network-analysis-and-the-government-fraudster/ (Accessed: 23 March 2026).

View source News article / media

Whittier Daily News (2009) 'Data mining catches welfare cheats', Whittier Daily News, 27 December. Available at: https://www.whittierdailynews.com/2009/12/27/data-mining-catches-welfare-cheats/ (Accessed: 23 March 2026).

View source News article / media
Deployment Status How far the system has progressed into real-world operational use, from concept/exploration through to scaled and institutionalised. View in glossary Full Production Deployment
Year Initiated The year the AI system was first initiated or development began. 2008
Scale / Coverage The scale and geographic or population coverage of the deployment. Los Angeles County CalWORKs Stage 1 Child Care Program; public reporting describes approximately 10-15 referrals per month during the documented operational period
Funding Source The source(s) of funding for the AI system development and deployment. Los Angeles County general funds and California state funding ($10 million state allocation for IHSS fraud prevention in October 2009)
Technical Partners External technology vendors, academic partners, or development partners involved. SAS Institute Inc. – SAS Fraud Framework for Government, including data mining, social network analysis, predictive analytics, rules management and forecasting components.
Outcomes / Results Public reporting on the pilot and early operational period states that the system identified collusive fraud rings with an 85 percent hit rate and generated an estimated $6.8 million in cost avoidance. In its first ten months of live use, DMS reportedly generated 197 additional child-care fraud referrals and 67 additional non-child-care referrals. Trade and vendor-adjacent reporting also states that the system surfaced more than 200 referrals overall and that roughly 10-15 cases per month were being referred during the documented period, with over 40 percent of referrals proving positive for fraud.

How to Cite

DCI AI Hub (2026). 'Data Mining Solution (DMS) for CalWORKs Stage 1 Child Care Program Fraud Detection', AI Hub AI Tracker, case USA-006. Digital Convergence Initiative. Available at: https://socialprotectionai.org/use-case/USA-006 [Accessed: 1 April 2026].

Change History

Created 30 Mar 2026, 08:42
by v2-import (import)