Calculated Risks: The Power of Predictive Risk Models to Regulate Families of Color

Author

• Ashleigh Washington, CUNY Graduate CenterFollow

Date of Degree

2-2026

Document Type

Doctoral Dissertation

Degree Name

Doctor of Philosophy

Program

Social Welfare

Advisor

Tricia Stephens

Committee Members

Nancy Giunta

Alexis Kuerbis

Subject Categories

Science and Technology Policy | Science and Technology Studies | Social Policy | Social Work

Keywords

Child Welfare, Family Policing, Artificial Intelligence, Predictive Risk Models, Surveillance and Technology Studies, Critical Race Studies

Abstract

This dissertation examines predictive risk modelling (PRM) in New York City Administration for Children's Services (ACS), specifically exploring AI-enabled design features, practices, and encoded and racialized assumptions about risk in child welfare prevention services. Guided by Critical Race Theory and the Algorithmic Ecology framework, this study traces how PRM in ACS evolved from a service termination tool into a broader, largely internal surveillance apparatus that leverages families’ personal information, including health information, to assess risk of future maltreatment and to evaluate contracted service providers. Drawing on analysis of public records and in-depth interviews with Black parents, the findings describe how the PRM operates using data collected during intrusive and coercive investigations.

While framed as a tool for fairness, the findings show little evidence of transparency and accountability. Parents were largely unaware of the model’s existence, yet their personal information shaped its calculations through investigation history and worker-driven assessments. This dissertation highlights how AI-enabled algorithms shape child welfare as a family policy system. It calls for critical examination of ethical machine learning, abolition of mandated reporting as a data pipeline, information justice in social work, and reparative policies that redress the harms of family policing toward resources for community-defined care.

Recommended Citation

Washington, Ashleigh, "Calculated Risks: The Power of Predictive Risk Models to Regulate Families of Color" (2026). CUNY Academic Works.
https://academicworks.cuny.edu/gc_etds/6635