Tue, 03/02/2021
This post is a summary of Episode 10 of The Nebraska Governance & Technology Center’s Podcast Series, Tech Refactored. Host Gus Hurwitz was joined by Woodrow Hartzog, Professor of Law and Computer Science at Northeastern University School of Law; Elana Zeide, Professor at the University of Nebraska College of Law; and Danielle Conrad, Executive Director of the American Civil Liberties Union of Nebraska.
As facial recognition technology has made its way into an increasing number of areas of life and society, from signing into our iPhones to the ways in which law enforcement operate, public awareness and concern over the ways in which facial recognition technologies are used have grown. In response a number of state legislatures, including Nebraska’s unicameral, have introduced bills that would curtail the ways in which law enforcement and other entities can use facial recognition. In an exciting development for the NGTC, Conrad contacted the center to see if we might be interested in exploring the issues surrounding facial recognition in an episode of Tech Refactored, and we jumped at the chance. After Conrad framed the issue in terms of the bill (LB 1091) coming before the Nebraska unicameral, Hurwitz opened the discussion up to the panel for a broader discussion of the promises and pitfalls of facial surveillance.
For Conrad, the purpose of Nebraska’s proposed ban on the use of facial recognition technology by law enforcement (LB 1091) is to ensure that “Nebraska puts a ban or a pause in place on this type of mass surveillance tool being utilized by our law enforcement agencies until we have a more thorough and better understanding about how to catch the law and the technology up to each other.” She cited research demonstrating that these technologies have a higher probability of misidentifying individuals from certain communities, including women, people of color, and gender-nonconforming Americans.
Hurwitz asked the panel whether concerns over facial recognition technology are overstated, essentially a “knee jerk reaction against this new technology” and whether there really are “powerful and socially useful” applications of facial recognition. Hartzog, a long-standing critic of facial recognition technologies, acknowledged that facial recognition does offer some benefits that can broadly be separated into two categories: a) “modest benefits” (being able to board an airplane a few minutes sooner, or being able to access your iPhone without having to use your thumb), and b) “significant benefits” (finding missing persons, or being able to apprehend individuals who pose immediate threats to others). For Hartzog, the question is “are the advantages significant enough to take on the immense cost of, what I view as a potentially dangerous — very dangerous — technology.”
Hartzog didn’t mince words in his analysis. “I think that facial recognition is the most dangerous surveillance technology ever created.”
I think there’s actually a lot of danger in taking an approach that allows facial recognition systems to become entrenched and normalized because of the inevitability of their abuse. (...) I don’t think there’s any scenario under which we come out better ahead with facial recognition technology. So I have advocated, and others have as well, for outright bans on facial recognition.
(For more information on Hartzog’s views on facial recognition technology, take a look at his fascinating article The Inconsentability of Facial Surveillance.)
Zeide also pointed out that there are settings where an individual can theoretically refuse consent to be subject to facial recognition (by, for example, declining to work at a company that uses such technology, or declining to fly on an airplane), but in a practical sense, such a decision so curbs an individual’s meaningful participation in society that it renders concepts of consent effectively meaningless.
Over the last few years, concerns about facial recognition technology have found purchase in a number of states and localities across the United States, resulting in various reforms or outright bans, particularly with regard to educational institutions. Zeide noted that “many universities have stopped their plans to institute facial recognition on the basis of protests; notably UCLA. And after reports about putting facial recognition cameras in several schools in upstate New York, governor Cuomo issued a moratorium on any use of these technologies until 2022.”
Personally, perhaps the most thought-provoking moment in the episode came when Hurwitz asked the panel what policymakers should be thinking about when addressing facial recognition technology. I would have assumed that the panel’s responses would be procedural or context-based reforms to minimize the risks, but Hartzog suggested a much bolder approach:
For me, I think the takeaway point is that facial recognition is not an inevitability and it doesn’t have to be. People create technologies and we can shape the way in which those technologies are formed. But we don’t have to follow the old playbook of following certain kinds of procedures, but letting everything flow.
We should be considering the many ways in which people want to control other people because facial recognition is at its base a tool of control. (...)
My takeaway would be that some technologies overall are not worth the risk. And facial recognition technology is probably one of those.
---
Our panel had some suggestions regarding sources for individuals looking for more information on facial recognition and its regulation:
Elana Zeide: Electronic Frontier Foundation; Epic; the American Civil Liberties Union’s work on facial recognition in schools; Ban Facial Recognition
Woody Hartzog: Georgetown Law’s Perpetual Line-up Project, The Future of Privacy Forum’s Explainer on Understanding Facial Detection, Characterization, and Recognition Technologies and Privacy Principles for Racial Recognition Technology in Commercial Applications
Tags: Tech Refactored Review
