EPIC Obtains Documents About DC's Use of Automated 'Risk Scores' for Public Benefit Recipients

EPIC, through a freedom of information request, has obtained new records about the D.C. Department of Human Services’ use of automated systems to track and assign “risk score[s]” to recipients of public benefits. The documents show that DCDHS has contracted with Pondera, a Thomson Reuters subsidiary, for case management software and a tool known as “Fraudcaster.” Fraudcaster tracks location history and other information about people receiving public benefits, combining this information with “DHS data and pre-integrated third-party data sets” to yield supposed risk scores. Factors that may cause the system to label someone as riskier include “travel[ing] long distances to retailers” and “display[ing] suspect activity.” Thomson Reuters also offered a free trial of its CLEAR service to the DCDHS as an incentive to sign the Pondera contract quickly. CLEAR is “powered by billions of data points” and claims to “identif[y] potential concerns associated with people.” The system is used by Immigration & Customs Enforcement and other law enforcement agencies in the U.S. EPIC is pursuing more information about DCDHS’s use of Pondera systems and mapping out automated decision-making tools used in D.C. through the EPIC Scoring and Screening Project. EPIC advocates for algorithmic transparency and accountability, particularly for systems used to make high-impact decisions like public benefit determinations.

« EPIC Submits Feedback on NIST AI Risk Management Framework | Main | GAO Report Finds 10 Federal Agencies Plan to Expand Use of Facial Recognition Through 2023 »

Share this page:

Defend Privacy. Support EPIC.
US Needs a Data Protection Agency
2020 Election Security