The Georgetown Law Center on Privacy and Technology recently completed an extensive study on America’s surprisingly extensive facial recognition software, and concluded half of American adults have already been tagged by systems accessible to law enforcement.
The study is disturbingly entitled The Perpetual Line-Up: Unregulated Police Face Recognition in America. As the title points out, there has been very little regulation of this technology from a government keen to regulate just about everything else into bureaucratic paralysis.
The introduction compares facial recognition databases to a police line-up nobody can refuse to participate in:
This summer, the Government Accountability Office revealed that close to 64 million Americans do not have a say in the matter: 16 states let the FBI use face recognition technology to compare the faces of suspected criminals to their driver’s license and ID photos, creating a virtual line-up of their state residents. In this line-up, it’s not a human that points to the suspect—it’s an algorithm.
But the FBI is only part of the story. Across the country, state and local police departments are building their own face recognition systems, many of them more advanced than the FBI’s. We know very little about these systems. We don’t know how they impact privacy and civil liberties. We don’t know how they address accuracy problems. And we don’t know how any of these systems – local, state, or federal – affect racial and ethnic minorities.
As the authors concede, facial recognition has real benefits, and has been successfully employed to “catch violent criminals and fugitives” by law-enforcement officers who are “men and women of good faith.”
However, facial recognition technology has implications for “privacy, civil liberties, and civil rights” – which makes it very similar to another advanced technology, drone aircraft. In fact, the two issues are linked, since drone cameras could ultimately be part of facial recognition databases, a combination featured in many dystopian science-fiction nightmares over the past century.
Indeed, The Perpetual Line-Up acknowledges that real-time facial recognition systems “seem like science fiction,” but the technology is very real, and our government is keenly interested in exploring its capabilities.
As with drones, facial recognition creates an unblinking electronic eye, which does much of its work at blinding speed, with minimal human intervention. In theory, it’s not much different from experienced police officers spotting the subject of an all-points bulletin and making an arrest, but Information Age technology makes that process so fast and omniscient that it becomes frightening. Citizens may not enjoy a reminder of just how many electronic eyes are watching them at any given moment, especially in urban settings.
A major concern highlighted by the Georgetown report is that old-fashioned law enforcement databases were compiled from “criminal arrests or investigations,” meaning the people contained within those databases had already encountered law enforcement.
“By running face recognition searches against 16 states’ driver’s license photo databases, the FBI has built a biometric network that primarily includes law-abiding Americans. This is unprecedented and highly problematic,” the report cautions.
The study authors advocate a more highly-developed legal and regulatory environment for facial recognition software, not suppression of the technology:
No state has passed a law comprehensively regulating police face recognition. We are not aware of any agency that requires warrants for searches or limits them to serious crimes. This has consequences. The Maricopa County Sheriff’s Office enrolled all of Honduras’ driver’s licenses and mug shots into its database. The Pinellas County Sheriff’s Office system runs 8,000 monthly searches on the faces of seven million Florida drivers—without requiring that officers have even a reasonable suspicion before running a search. The county public defender reports that the Sheriff’s Office has never disclosed the use of face recognition in Brady evidence.
There is a real risk that police face recognition will be used to stifle free speech. There is also a history of FBI and police surveillance of civil rights protests. Of the 52 agencies that we found to use (or have used) face recognition, we found only one, the Ohio Bureau of Criminal Investigation, whose face recognition use policy expressly prohibits its officers from using face recognition to track individuals engaging in political, religious, or other protected free speech.
Inaccuracies in the system are one of the primary regulatory concerns. Ironically, the problem human observers have in verifying photo matches is one of the major accuracy concerns. This is especially true for African-American subjects, as a study co-authored by the FBI concluded “face recognition may be less accurate on black people,” while their high arrest rates suggest they will be the subjects of recognition scans more frequently.
It’s possible that facial recognition systems are more accurate than the paltry handful of extant studies suggest, or that they can be made far more accurate, but a major theme of The Perpetual Line-Up is that more research needs to be done, before the use of facial recognition software grows even more widespread. For that matter, the authors contend the public is not yet sufficiently aware that these systems exist.
The report’s recommendations will be familiar to every student of Surveillance State issues: there should be tighter controls over the use of facial recognition technology, proper legal warrants should be issued, the databases should be limited to law-enforcement images (i.e. mug shots) rather than driver’s license photos to keep people with no criminal record out of the system, the accuracy of these systems should be studied carefully and improved, with state and federal funding contingent on the maintenance of rigorous standards…
But won’t there be arguments from the security side of the spectrum that facial recognition could be invaluable for catching suspects with no previous criminal record?
Consider the recent example of Arcan Cetin, the Cascade Mall shooter, who was mistakenly identified as Hispanic based on a mall security camera image. Capturing this dangerous killer was a high priority. Identifying him more quickly would have been helpful. Advanced facial recognition systems might be able to do that, but only if they have the necessary data, and Cetin’s encounters with law enforcement might not have been serious enough to get his face into a strictly regulated database. This could be a recurring problem with recent immigrants, young native-born people radicalized into terrorism or crime, or others who commit extremely serious first offenses.
If the systems become highly accurate, and their use is carefully monitored from a legal standpoint, there will be great pressure to make the pool of available facial data as large as possible.
Also, if real-time facial recognition becomes extremely accurate, its commercial uses will doubtless increase, as perhaps foretold by the film Minority Report, which showed as a world full of automated advertising tailored on the fly to each instantly-identified passerby. (The movie’s futuristic tech used retinal scans, rather than facial biometrics.) Audiences in 2002 laughed, a bit nervously; a young person watching the film today might see those pervasive automated advertisements as an extension of what Facebook and Google already do.
Consumer acceptance of such technology, or perhaps a broad generational shift in which young people shrug and assume their every step is already tracked in a thousand different ways, could break down resistance to merging security and commercial databases. A poll of whether people in various age groups are troubled by the fact that half of American adults are already in facial recognition databases would be interesting. The authors of The Perpetual Line-Up are right to call for solid legislation and regulatory practices before this genie gets any further out of the bottle, into an America where the air is already thick with digital genies.