US seeks input for biometrics equity study to determine course of remote government access
The U.S. government is asking for contributions to a study into equity in remote identity proofing for Login.gov, attempting to determine how to best meet the NIST SP 800-63-3 Identity Assurance Level (IAL) 2 standard without introducing bias through biometric algorithms.
The General Services Administration’s (GSA’s) Technology Transformation Services (TTS) department is carrying out the study, as announced by GSA Executive Director of Identity Phil Lam.
IAL2 is the highest degree of certainty provided in remote identity proofing, and therefore is specified by NIST to require biometrics, if not carried out in person.
“This data study will enable GSA to make a data driven decision on whether to pursue facial verification capabilities for Login.gov in the near term, to determine baseline performance metrics, and to provide real-world identity verification pass rate data for the first time to the broader Federal agency community,” the study document states.
The Source Sought document outlines a plan to produce a peer-reviewed report in 2023 to assist decision-making regarding digital identity verification.
“This remote identity proofing study will also assess vendor performance on the non-biometric aspects of the NIST SP 800-63-3 Identity Assurance Level (IAL) 2 flow,” the document further explains. “This will allow for a more informed comparison on drop-off rates, i.e., to what extent a tester passes or fails the flow is caused by the biometric versus the non-biometric aspects of the IAL2 proofing process.”
Specifically, the GSA TTS Equity Study on Remote Identity Proofing seeks recruiting services to bring in 2,000 testers, build out and provide a dashboard for metrics, and conduct field tests. It also seeks participant outreach services to develop the strategy and partner with the recruiter “to reach target demographic groups,” and a spoof testing service to conduct negative testing with a minimum of 500 transactions, and then “(d)raft and deliver a final report and dashboard with metrics to compare against vendor results.”
Both the recruiting service and outreach partner are expected to specialize in biometrics or machine learning testing.
The study follows the release of an ‘Equity Action Plan’ that revealed findings that people with dark skin using face biometrics to access government services met with higher false rejection rates than other groups.
Responses are sought by May 2, 2022.