Civil Rights Commissioner Mondaire Jones says that he's taken a "dim view" of the lack of cooperation by two federal agencies with AI oversight.

Civil Rights Commissioner Mondaire Jones says that he's taken a "dim view" of the lack of cooperation by two federal agencies with AI oversight. Eugene Gologursky/Getty Images

Civil Rights commission digs into government use of facial recognition

The tech poses “serious threats to our fundamental rights,” the chair of the commission said during a Friday briefing.

The U.S. Commission on Civil Rights — an independent agency meant to inform national civil rights policy and its enforcement — is working on a report on the federal government’s use of facial recognition. 

The commission, composed of bipartisan members appointed by the President and Congress, held a briefing on Friday to gather ideas for the forthcoming report. Commissioners asked about how facial recognition technology is trained and tested, how well it works across different populations and how the tech is used by government agencies and law enforcement.

“While the technology offers potential benefits, it also possesses serious threats to our fundamental rights,” said Rochelle Garza, the commission’s chair, during the briefing. “One of the most pressing concerns of [facial recognition technology] is its disproportionate impact on marginalized groups… and beyond the issue of bias, [facial recognition technology] poses a stark threat to our constitutional freedoms.”

The commission heard from representatives across the Department of Homeland Security, White House, civil society groups, vendors and more, although commissioners expressed frustration at the lack of participation from the Departments of Justice and Housing and Urban Development.

“I've had to take a dim view of why these two departments have chosen not to cooperate with the commission's legitimate inquiry into their use of facial recognition technology, which suggests to me that DOJ and HUD are embarrassed by their failures and are seeking to avoid public accountability,” said Mondaire Jones, the commissioner leading the effort and a former congressman for the state of New York.

A Justice spokesperson told Nextgov/FCW that the department’s written submission “will be forthcoming soon.” 

HUD told Nextgov/FCW that it “is cooperating with the U.S. Commission on Civil Rights and provided answers to the Commission’s extensive interrogatories and document request in advance of the Commission’s briefing on facial recognition technology” and “plans to submit written testimony,” noting that the department doesn’t use the tech but “urges its program participants to find the right balance between addressing security concerns and respecting residents’ right to privacy.”

Washington Post reporting referenced during the briefing found that HUD grants had contributed to the proliferation of facial recognition tech in public housing. 

Across the federal government, the Government Accountability Office reported in 2021 that 18 of 24 surveyed federal agencies were using facial recognition, with common uses being for law enforcement and digital access, including unlocking agency cellphones. 

In the law enforcement context, the tech has the potential to help solve crimes and identify unknown suspects, but it also offers the potential for error that could lead to the misidentification, arrest and prosecution of an innocent person, said Gretta Goodwin, director of Homeland Security and Justice at GAO, during the briefing. 

Officials from DHS, which issued a policy on the department’s use of facial recognition tech in September, told commissioners about use cases at airports and ports of entry. To date, Customs and Border Patrol has used the tech to identify over 1,900 imposters trying to enter the U.S., said Diane Sabatino, acting executive assistant commissioner at CBP’s Office of Field Operations. 

But concerns remain. The Government Accountability Office found in a 2023 report that seven law enforcement agencies used the tech without requiring training of staff beforehand, said Goodwin. Whether or not federal law enforcement agencies had specific guidance or policy on facial recognition varied.

The National Academies of Sciences also issued a report earlier this year recommending that the government act to regulate facial recognition — something commissioners appeared to be paying attention to as well.

“I continue to be very concerned by the lack of guardrails at the federal level when it comes to the use of facial recognition technology,” Jones told Nextgov/FCW on the sidelines of the briefing. 

Many skeptical of the technology during the briefing pointed to 2019 testing by the National Institute of Standards and Technology showing differentials in how well facial recognition algorithms work depending on factors like race and gender. NIST officials have since said that the tech has generally improved since then, however, and that how well it works varies across vendors.

Deirdre Mulligan — principal deputy U.S. chief technology officer at White House’s Office of Science and Technology Policy — told commissioners that the draft guidance on artificial intelligence released by the Office of Management and Budget last year “would establish a rigorous set of risk management processes and requirements for government use of rights impacting AI, including [facial recognition technology].”

Mulligan also pointed to ongoing interagency work to develop a report with recommended guidelines for law enforcement agencies, per a 2022 executive order — a topic area of concern for some witnesses and commissioners.

The commission’s report will be published Sept. 30, 2024, according to Angelia Rorison, the commission’s media and communications director. The commission is taking additional materials from the public through April 8.