Privacy Implications of Facial Recognition

Demystifying Clearview AI Blog Series (Part 9)

6 min readDec 30, 2020

Photos by Drew Angerer, *Sam Simmonds, and* *Kena Betancur.*

“The flames kept building until both cabinets were empty
Woody insisted on sitting in the garage until all that remained was ash.
Lance could only guess at what he’d helped to destroy.”
- The Secret History of Facial Recognition, Shaun Ravin

What He’d Helped to Destroy

Throughout this series, we've learned how facial recognition technology is flawed by design; ominously, those flaws are masked behind statements such as “independently tested…99% accuracy.”

While the technology itself is neither good, nor bad, nor neutral, law enforcement’s application deserves conscious scrutiny because it implicates important Constitutional values. We may very well decide that values such as privacy and the right to be free from unlawful searches and seizures are overrated, in which case we should have the appropriate John Hancock step forward such that posterity may know what generation’s wisdom to be forever thankful for.

As explained by Jennifer Lynch of the Electronic Frontier Foundation during her testimony to the Senate Subcommittee on Privacy, indiscriminate use of facial recognition alters the traditional presumption of innocence by placing more of a burden on an individual to show he or she is not who the system identifies him or her to be. That’s one of the many reasons the EFF is urging courts to mandate that agencies disclose the facial recognition algorithms and models used when charging someone with a crime. There is no such known requirement at any state or federal level.

Abuse of Due Process and Erosion of Equal Protection

Facial recognition technology has been in use for decades. There’s no question it’s effective. Touting the successes of the technology doesn’t negate its consequences. Without proper safeguards, law enforcement’s use of facial recognition amounts to a backdoor repeal of Due Process and a wanton rejection of Equal Protection under law.

Facial recognition as an attack on First Amendment activities isn’t a hypothetical; it is already a reality. In 2015 Baltimore PD partnered with the location-based analytics platform Geofeedia to track protesters at a peaceful protest. During the protest, police ran facial recognition to find individuals with “outstanding warrant and arrest them directly from the crowd.” The openly admitted goal of those arrests was not enforcing warrants but rather dissuading participation in the protest.

For more on the implications of facial recognition, such as the identity theft dangers of standardizing on faces as an identifier, see Jennifer Lynch’s testimony to the Senate Subcommittee on Privacy. Facing the Future of Surveillance by the Constitution Project’s Task Force on Facial Recognition Surveillance is another excellent resource laying out the key areas of concern and policy recommendations for properly embracing facial recognition in a manner that both protects constitutional principles and aid public safety.

Fake News: Your Privacy is Not At Risk

We introduced this series with a short history of facial recognition; we’ll end with a much longer history of facial recognition. A history that starts with the 121-year old technology giant you’ve never heard of, NEC.

NEC’s biometric contracts in the US. Source: NEC

Formerly known as the Nippon Electric Company, NEC started its fingerprint identification business unit in 1969, working with the Japanese National Police Agency. By the 1980s, it had expanded internationally with clients in the US and across the globe. It’s research and development of commercial facial recognition technology started in 1989. NeoFace, NEC’s first mass-market facial recognition software, was released in 2002. Currently, NEC has thousands of active biometrics contracts with agencies across the US and around the world.

NEC’s History of Biometrics Technologies development. Source: NEC

Broad market consumer facial recognition technology in the US dates back to 2004 with the DARPA funded highly effective PittPatt startup, purchased by Google in 2011. Firms such as Tapwire, Terrogence, Verint, WOLFCOM, and banjo have been in business with law enforcement agencies for much longer than Clearview. With banjo especially having a reputation and methods even more notorious than Clearview. According to a 2016 study by the Center on Privacy and Technology at Georgetown Law, at least one in four state or local police departments have the ability to run facial recognition searches either directly or via a partnering agency.

Facial recognition has been in use by local, state, and federal government agencies for decades. It has never implied such a severe threat to privacy primarily because industry contained its use to criminal databases, themselves subject to GDPA and CCPA type of disclosure and retention requirements.

Real News: Your Privacy is Endangered!

Scenes from Person of Interest (TV Series)

What Clearview represents is the threat of legitimizing the “unchecked infrastructural expansion” of a powerful technology.

As further explained by Jennifer Lee of the ACLU, this technology stands to be used to deny people essential services such as access to housing, health care, food, and water. Stories of false arrests by police due to faulty facial recognition are well known. Yet, even beyond that, a mistaken identity during a background check could cause individuals to be denied employment or promotions without recourse. Given the disproportionate error rates for women and people of color, unchecked use of this technology risks magnifying systemic discrimination to a new level.

Privacy has been dying for a very long time; Clearview is the stench of corpses impossible to ignore. Forcing us to address whether we want to bury it or burn it.

First they came…

For years the ACLU and over a hundred organizations in the U.S. and worldwide have warned of the dangers facial recognition poses to vulnerable groups. In the beginning, arguments focused on the accuracy of the technology; the response has been the rollout of more accurate technology. As the technology has become easier to use, we’re now seeing it deployed on its most ardent supporters.