Facial recognition technology helps nab criminals—and raises privacy concerns
Accuracy levels depend on myriad factors, including the conditions under which the photo was taken. In general, the technology has a high accuracy rate for posed photos, such as passport pictures, in which people are told to look directly at the camera, says Anil Jain, a computer science professor at Michigan State University.
But the technology is less accurate for photos taken under noncontrolled conditions, such as when street cameras capture images as subjects are looking away, wearing caps or making expressions that distort their features.
No federal laws currently constrain the use of facial recognition software, and most observers think national legislation is unlikely, given the divided political environment.
But on the state level, Illinois, Texas and Washington have laws that aim to protect people’s biometric privacy in the commercial sphere. Other states, including Alaska, Connecticut, Montana and New Hampshire, are considering regulating the use of the facial recognition technology.
The Illinois Biometric Information Privacy Act, which was passed in 2008, prohibits private sector companies from collecting some kinds of biometric identifiers, including scans of facial geometry, without obtaining people’s written consent.
That law has already sparked several lawsuits, including putative class actions against Facebook, Google and Shutterfly. The plaintiffs in those cases allege that the tech companies illegally created “faceprint” databases. Google declined to comment, and Facebook and Shutterfly didn’t respond to interview requests.
Shutterfly settled one lawsuit in spring 2016 and was sued a second time last fall. Facebook and Google have been fighting the allegations in court since 2015.
The companies raise several arguments, including that the Illinois statute is ambiguous about whether it applies to photos. The law says the restrictions don’t apply to photos or information derived from them, but the statute also appears to be aimed at regulating the technology.
Attorney Jeffrey Neuburger, co-chair of the technology, media and telecommunications group at Proskauer Rose, says the Illinois law was written before biometrics were prevalent, and that there’s very little legislative history to shed light on why photographs were excluded from the law’s definitions.
The plaintiffs argue that the law wouldn’t make sense unless it covers face scans derived from photos. So far, the plaintiffs have prevailed with that argument. “Courts have said excluding that kind of information undermines the purpose and effect of the statute,” Neuburger says.
Those lawsuits also pose questions about whether states can regulate the use of facial recognition software by a national company.
RISK OF IDENTITY THEFT
Jay Edelson, a plaintiffs lawyer in Chicago who is involved in the Facebook case, says that biometric databases could expose people to the risk of fraud.
Databases of fingerprints, voiceprints and faceprints will “be a major target for identity thieves and other bad actors out there,” he says.
Edelson speculates that fraudsters could use a 3D printer to create a mask based on a faceprint, and that it could fool cameras.
But others say there’s no call for laws that could restrict companies from rolling out new services that at least some consumers seem to like—such as Facebook’s tagging feature, which can use facial recognition technology to help people automatically attach their friends’ names to photos.
“Identity theft is illegal today, whether it’s done by impersonating me or by going to a photo database, downloading a model of my face, and printing a 3D mask like in Mission Impossible,” says Carl Szabo, senior policy counsel at NetChoice, a trade group that represents internet companies.
Szabo points out that the technology can have “a multitude of potential positive uses,” including improved security at banks and other establishments.
“Where we see abuses, we should target those abuses,” he says. “What we need to make sure we avoid is the privacy panic.”
This article appeared in the October 2017 issue of the ABA Journal with the headline “Face Time: Facial recognition technology helps nab criminals—and raises privacy concerns."