From the U.S. Government Accountability Office, www.gao.gov Transcript for: Federal Law Enforcement’s Use of Facial Recognition Technology Description: The use of facial recognition technology is expanding, including among law enforcement. While its use may help law enforcement solve crimes more quickly, there are concerns about the extent to which federal agencies are monitoring their own use of facial recognition technology. We talk with GAO’s Gretta Goodwin to find out more. Related GAO Work: GAO-21-518, Facial Recognition Technology: Federal Law Enforcement Agencies Should Better Assess Privacy and Other Risks Released: June 2021 [Intro music:] [Gretta Goodwin:] There are innumerable ways in which the technology might be abused. Risk still exists that searches will provide inaccurate results. [Holly Hobbs:] Hi and welcome to GAO’s Watchdog Report. Your source for news and information from the U.S. Government Accountability Office — celebrating 100 years of fact-based, nonpartisan government oversight. I'm Holly Hobbs. The use of facial recognition technology is expanding, including among law enforcement. While its use may help law enforcement solve crimes more quickly, there are concerns about the extent to which federal agencies are monitoring their own use of facial recognition technology. Today we'll talk with Gretta Goodwin, an expert on federal law enforcement and a director in our Homeland Security and Justice Team, about her new report on federal law enforcement's use of facial recognition technology. Thank you for joining us Gretta. [Gretta Goodwin:] It's a pleasure to be here, Holly. [Holly Hobbs:] So Gretta, can you give us some examples of how federal law enforcement agencies are using facial recognition technology? [Gretta Goodwin:] Sure, federal law enforcement agencies use the technology for a number of reasons, ranging from supporting criminal investigations, to surveillance, to using it to meet program requirements. A federal investigator might have a photo of a person suspected of committing a crime, but they may not know the identity of the person in the image. So the investigator can use the technology to compare the photo to a gallery of photos, like mugshots or drivers’ licenses. In the case of surveillance in considering the security operations at the White House complex, the Secret Service piloted a system with facial recognition technology. While they ultimately decided not to implement the system, it had been considered as a tool of surveillance. And then, to help meet COVID-19 safety measures, the Administrative Office of the U.S. courts began using facial recognition technology to meet requirements in a socially distance manner. For example, they allowed individuals under court-ordered supervision to verify their identity via a smartphone app rather than meeting with their probation officer. [Holly Hobbs:] And do we know how often or how broadly federal law enforcement is using this technology? [Gretta Goodwin:] So, we surveyed 42 federal agencies that employ law enforcement officers about their use of facial recognition technology. Twenty of those agencies reported owning systems with the technology or using systems owned by other entities like state, local, or non-government organizations. [Holly Hobbs:] So are the federal agencies monitoring who is using or has access to it, and how are they ensuring it's not being abused? [Gretta Goodwin:] So not always. 14 of the federal agencies we reviewed told us that they used the systems to support criminal investigations. We found that only one of those 14 agencies had awareness of which non-federal systems were being used by their employees. The remaining 13 did not. When we asked one of the agencies about its use of non-federal systems, officials told us that they had to poll their employees because the information was not maintained by the agency. In another instance, officials from an agency initially told us that their employees did not use non-federal systems. But, after they conducted a poll, they learned that the employees had used the system to conduct more than 1,000 facial recognition searches. [Holly Hobbs:] Can you give us an example of how law enforcement might abuse this technology? [Gretta Goodwin:] Holly, there are innumerable ways in which the technology might be abused. There is a risk that federal agencies do not adhere to privacy-related laws. And although the accuracy of facial recognition technology has increased dramatically in recent years, risk still exists that searches will provide inaccurate results. For example, it could identify innocent people as investigative leads. [Holly Hobbs:] And I know GAO is working on a report about the January 6 attacks on the US Capitol. And that that will be released at a later time. But do we know whether and how facial recognition technology was used to ID people involved in that attack? [Gretta Goodwin:] We do. So three agencies reported using the technology on images of the Capitol attack. These agencies are the Capitol Police, U.S. Customs and Border Protection, and the State Department's Bureau of Diplomatic Security. However, additional agencies might have used the technology. But to give you a sense for how the technology was used, the Capitol Police told us that they used face recognition services provided by a private company to help identify individuals in the images of the Capitol attack. U.S. Customs and Border Protection and the Bureau of Diplomatic Security told us that they performed facial recognition searches at the request of another federal agency. Now in addition to the Capitol attack, 6 agencies reported using the facial recognition technology on images of the civil unrest, riots, or protests following the death of Mr. George Floyd in May of 2020. For example, the Postal Inspection Service used the technology to help identify individuals suspected of stealing mail, burglarizing the Postal Service buildings, or committing arson. And then the FBI used the system to help identify people in the images they received via the tip line. [Music] [Holly Hobbs:] So it sounds like facial recognition technology is a powerful tool that can be used to help generate investigative leads and solve crimes, but that federal law enforcement agencies do not monitor its use by employees to ensure it's being used appropriately. Gretta, did we make any recommendations that would improve the oversight of this technology? [Gretta Goodwin:] We made recommendations to 13 federal agencies that each of these agencies take steps to understand what non-federal systems are being used by their employees for criminal investigations. And that they assess the risk of using these systems, including the privacy and accuracy related risk. [Holly Hobbs:] And last question. What's the bottom line of this report? [Gretta Goodwin:] Holly, we found that 14 federal agencies use facial recognition systems to support criminal investigations, and all 14 use systems owned by non-federal entities like state police or private companies. So it is essential that federal agencies have thorough understanding of what systems are being used by their employees. They also need to ensure that employee use of these systems complies with privacy laws and that these systems are sufficiently accurate. [Holly Hobbs:] That was Gretta Goodwin talking about GAO's recent review of the use of facial recognition technologies by federal law enforcement. Thank you for your time, Gretta. [Gretta Goodwin:] Thank you Holly. [Holly Hobbs:] And thank you for listening to the Watchdog Report. To hear more podcasts, subscribe to us on Apple podcasts. And make sure you leave a rating and review to let others know about the work we're doing. For more from the congressional watchdog, the U.S. Government Accountability Office, visit us at GAO.gov.