Most US government agencies use facial recognition

A new report from the Central Government Accountability Office (GAO) found that 19 out of 24 US government agencies surveyed use facial recognition in some way, illustrating how commonplace the controversial technology has become within the federal government. The list of agencies includes agencies such as the Department of Defense and the Department of Homeland Security (DHS) that maintain internal systems, as well as smaller agencies that use the system to control access to highly secured locations.

“As the use of FRT continues to grow,” the report explains, “members of Congress, academics and advocacy groups have emphasized the importance of developing a comprehensive understanding of how it is used by federal agencies.”

An earlier GAO report found that federal agencies have few meaningful systems to track the use of privately contracted facial recognition systems.

A graph showing the results of the GAO survey.

One system described in the report, called “TacID Guard Dog,” is used by the Department of Energy to “monitor the entry and exit of controlled locations.” The department bought the system in December 2019 and spent about $150,000 to test it over the course of the following year. The same system is also used by the Ministry of Defense.

The federal government operates a number of facial recognition systems. DHS uses the Automated Biometric Identification System to process travelers at border crossings, while the FBI’s Facial Analysis, Comparison and Evaluation service (or FACE) provides broad matching capabilities in criminal investigations.

The report found four separate agencies that had used Clearview AI, a controversial private company that compares new photos with data scraped from social media platforms. DHS and the Department of Justice used Clearview AI for conventional police searches, while the Department of the Interior used the system through the US Park Police. Notably, the Inspector General’s office at HHS also began piloting Clearview AI in September to “help identify subjects of criminal investigation.”

Privacy advocates have expressed deep concern about racial bias in facial recognition and states like Maine and Massachusetts have taken specific measures to limit their use. However, those efforts have done little to slow down federal use of the technology. In particular, Customs and Border Protection continues to expand the use of facial recognition at airports, both as part of the Biometric Exit program and in voluntary programs such as Global Entry. The TSA is currently testing a new facial recognition system as part of its credential verification system, though it remains unclear whether the system will roll out to wider use.