The Politics of Facial Recognition Systems: Issues and Policy
Facial recognition (FR) systems are considered a promising addition to the arsenal of technologies to be deployed in a battle against crime and terror. As a tool for scanning large numbers of people they would greatly improve the ability to identify known criminals and terrorists and even, in certain locations, function as a deterrent. Although there has been steady progress in the science and technology of video surveillance and FR systems, they are still far from perfect. The focus of our proposed research project is a particular imperfection in FR systems that may have significant political ramifications.
A study of commercially availably FR systems, which evaluates them against a comprehensive range of performance criteria, finds that most recognition algorithms, much like human recognition heuristics, are biased. As such they are better at recognising minorities within a population and, likewise, show biased results in relation to gender, age and ethnicity. These biases become even more significant if tied to known biased practices such as the targeting of minorities by CCTV operators. Furthermore, it has been shown that although FR systems are very good at performing identification tasks with small populations in controlled settings, their performance tends to degenerate quite rapidly as the population size increases and the conditions within which the systems operate deteriorate; in such conditions, their biases become amplified.
- Complete a report “Policy and Implementation Issues of Facial Recognition Systems” for distribution to informed policy makers and senior law enforcement officials with regard to the purchase, implementation, and use of facial recognition systems for law enforcement and security purposes. It will provide an overview of “state of the art” as well as implementation and policy challenges.
- Host a 5-day visit by Lucas Introna, Lancaster University UK, collaborator on the project.
- Host a talk by Dr. Jonathan Phillips, NIST. Phillips heads the Bureau’s efforts to monitor state of the art of Facial Recognition systems through the Facial Recognition Grand Challenge (latest one issued Sept 30, 2005) and Facial Recognition Vendor Tests (new round beginning Jan 30, 2006).
About Helen Nissenbaum, Project Director
Helen is an Associate Professor of Culture and Communication and Computer Scienc at NYU, as well as a Senior Fellow of the Information Law Institute at NYU School of Law. She is currently the co-editor of the Journal of Ethics and Information Technology. Helen is the recipient of numerous grants from the National Science Foundation, as well as from the Ford Foundation. She has previously served as the Associate Director of the University Center for Human Values and the Program in Ethics & Public Affairs at Princeton University, and is the author or editor of several books, including Emotion and Focus and Academy and the Internet.
Her specialties include philosophy and politics of technology: ethical, social and political implications of information and communications technology (privacy, security, accountability, intellectual property rights, electronic publication, computing in education); Information and communications policy; Applied and professional ethics (engineering ethics, scientific integrity and research ethics).
Helen received her Ph.D. in Philosophy and M.A. in Social Sciences in Education from Princeton University and her B.A. from the University of Witwatersrand, South Africa.
Departments of Culture & Communication and Computer Science
239 Greene Street, 7th Floor
New York, NY, 10003