Today, the ACLU of Massachusetts and two technology experts affiliated with the Berkman Klein Center and MIT Media Lab filed a Freedom of Information Act request with the Department of Homeland Security to learn about the federal government’s use of facial recognition technologies. The ACLU and Boston-area technologists, represented by the firm Fish & Richardson, seek information about how Customs Border Protection (CBP) and the Transportation Security Administration (TSA) use facial recognition in locations throughout the country, including at airports and other border crossings. Government disclosures and press reports confirm that these agencies are using facial recognition technologies, but the details about these programs are shrouded in secrecy, and DHS has yet to formalize rules to govern its biometric exit program.
During Trump’s presidency, the use of biometric exit face scans at airports has increased to nine U.S. international airports. Last year, CBP and Jet Blue announced a pilot program in Boston, whereby passengers on flights to Aruba could use their faces instead of tickets during the boarding process. Also in 2017, TSA announced a plan to use facial recognition at airport security checkpoints.
Despite the serious privacy and civil liberties implications of existing government facial recognition deployments, the Department of Homeland Security has not published rules to control how the technology is used, or how the data the government collects and accesses will be retained or shared. Last week, Senators Edward Markey (D-MA) and Mike Lee (R-UT) sent the Department of Homeland Security a letter urging them to “formalize its process and procedures for collecting passenger data before expanding the program.” The Senators also asked DHS for information about the process for opting-out of the face scanning system, if such a process exists.
The ACLU records request, filed in cooperation with technologists Thom Miano and Daniel Pedraza, asks for information about the policy framework governing DHS’ facial recognition systems, as well as information about the underlying technology. Specifically, the request seeks to uncover which companies or individual machine learning experts designed the algorithms that animate the facial recognition systems in use by DHS; any audits reviewing the performance of the algorithms; information about the systems’ accuracy rates, including any records describing accuracy rates by gender and race; policy directives, memoranda, and contracts related to the facial recognition technology; and all documents describing the legal authorization for DHS face scanning programs.
Mr. Miano and Mr. Pedraza are collaborators in the 2018 cohort of the Assembly program, run by the Berkman Klein Center at Harvard Law School and the MIT Media Lab. The Assembly program is a hands-on, experimental research arm of the Ethics and Governance in Artificial Intelligence Initiative, a joint endeavor of Berkman Klein and the Media Lab.
Read the FOIA request.