The FBI came under fire at a June 4 House Oversight and Reform Committee hearing for allegedly violating policies governing facial recognition systems, as lawmakers explore potential legislation to regulate the technology.
Rep. Jim Jordan (R-Ohio), the panel’s ranking member, pointed to a 2017 Government Accountability Office report that the FBI hasn’t complied with privacy and accuracy recommendations for its facial recognition program. The GAO said the FBI didn’t publish a timely Privacy Impact Assessment as required and didn’t file a system of records notice, which describes how information is used and protected, until after the GAO completed its review.
“How are we supposed to have confidence in strict policies that you’re going to follow when you didn’t follow the rules when you set up the thing in the first place?” Jordan said to Kimberly Del Greco, the FBI deputy assistant director of the Criminal Justice Information Services division. “We’re supposed to believe ‘Don’t worry, everything’s just fine?’”
Del Greco said the FBI follows “strict policies,” and the Department of Justice disagrees with the GAO’s assessment of the legal requirements. “We do care about the accuracy of the system and the testing,” Del Greco said.
The FBI’s facial recognition program doesn’t directly identify subjects, Del Greco said. It produces anywhere between two and 50 possible matches for investigative law enforcement purposes, she said. The technology also isn’t designed for real-time capture; it instead compares photos against a database of arrest mugshots and state driver’s license databases, she said.
The program includes the ability to submit a photo for a suggested match to 21 state motor vehicle department drivers’ license databases. Some lawmakers voiced concern that elected officials were excluded from the decision to authorize the FBI’s use of state DMV databases.
“Do any elected officials have anything to do with these decisions?” Committee Chairman Elijah Cummings (D-Md.) asked. “Where is that authority coming from? We’re trying to figure out with something affecting so many citizens, whether elected officials have anything to do with it.”
The hearing also featured testimony from the GAO, Transportation Security Administration, and the National Institute of Standards and Technology about the TSA’s facial recognition pilot programs. Lawmakers questioned the accuracy of algorithms, which critics have faulted for being unable to distinguish differences among races.
The TSA is piloting one facial recognition program at a baggage check in Atlanta’s international airport. Passengers can opt out of the program, which takes a picture of the person checking the bag, Austin Gould, a TSA assistant administrator for requirements and capabilities analysis, said.
“There was no automatic capture of passengers,” Gould said, adding that the system doesn’t retain images following a successful match.
Jordan and Cummings expressed interest in continuing a bipartisan investigation into government use of facial recognition technology and have previously expressed interest in bipartisan legislation.
Cummings said after hearing that he will call back GAO and FBI witnesses later this summer and plans to ask some members to take a “deep dive” into facial recognition technology.
“I know we have to balance law enforcement’s desire to have another tool in their toolbox, but we’ve got to make sure that that tool is not defective,” Cummings said after the hearing. “I am concerned that it is a not a program that’s ready for prime time and we’ve just got to look into it very carefully.”
—With assistance from Rebecca Kern