The Governmental Accountability Office (GAO) has released an extensive report on the facial recognition habits of US government agencies. The report includes feedback from 24 organizations, 19 of which are using facial recognition technology and 10 of which are actively researching new applications of the tech internally.
Some of those applications are relatively uncontroversial. For example, the Department of Transportation is using facial recognition to monitor tired and distracted drivers, delivering a safety feature that is becoming increasingly popular with commercial vehicle manufacturers.
However, the government’s enthusiasm for facial recognition could prove to be far more contentious. Many departments are trying to improve the accuracy of facial recognition technology more generally, with some focusing on the effects of aging and others trying to identify people who are wearing masks. That may not be a problem in the abstract, but it could raise concerns with privacy advocates if deployed in service of mass surveillance operations.
In that regard, the GAO report found that access control and law enforcement were the two most common use cases for facial recognition. The former is becoming more and more common in both the public and private sectors, and can secure a facility or a digital application while still preserving people’s privacy.
The same is not necessarily true of law enforcement, especially since the GAO report suggests that the government is being extremely cavalier with the technology. Many of the agencies that are using facial recognition have contracted with private developers, but 13 of those agencies had not familiarized themselves with those private solutions and had not yet performed a full privacy and accuracy assessment. Several agencies – including the Justice Department, the Air Force, the Fish and Wildlife Service, and Immigration and Customs Enforcement – had also partnered with Clearview AI, which is currently embroiled in numerous legal conflicts stemming from its flagrant disregard for consent and privacy laws.
To make matters worse, several other agencies acknowledged that they had performed Clearview searches even though they hadn’t entered into a formal agreement with the company. In those cases, federal agents had asked state and local agencies to perform the search and share the results.
All told, ten federal agencies reported that they want to increase their use of facial recognition ahead of 2023. The full list includes the Departments of Agriculture, Commerce, Defense, Homeland Security, Health and Human Services, Interior, Justice, State, Treasury and Veterans Affairs. That push to expand the government’s facial recognition capabilities comes amidst growing backlash to the technology, and as more state and municipal governments try to ban its public use. Federal lawmakers have also introduced similar legislation in Congress.
Unfortunately, the government’s continued use of Clearview AI (both on and off the books) would seem to indicate that it does not put much stock in public opinion, and may even try to circumnavigate any bans that do get put in place. Having said that, the GAO has argued that Congress should implement stronger facial recognition standards, and that federal agencies should be more transparent with their use of the technology.
August 27, 2021 – by Eric Weiss