Facial Recognition Surveillance at the Library?
By: Allyson Mower
A new book by Christopher Andrews called The Secret World is set to be released in the U.S. next week. It’s about the history of intelligence and surveillance. I’m reading an advanced copy as part of my ALA member benefit with NetGalley and as a way to continue learning about privacy. I’m a few chapters in and so far have learned that, historically, surveillance was used by political states on those who wished to subvert the political system and by armies during times of war. The author has not yet provided an example of a past society surveilling for the purpose of deterring potential criminals. That seems to be a concept only modern day societies have invented. We live in a day and age when camera surveillance has become ubiquitous as a way to deter crime (other purposes include prosecution of criminals and to make people feel safe).
As a case in point, the Van Houten Library at the New Jersey Institute of Technology (NJIT) implemented face detection software in February 2018 for precisely that purpose. Shydale James reports that in 2014 eleven property thefts had occurred at the library and the director of security sought to reduce that number, which, according to the report, has been achieved since installing the system. The system, called FacePro, was developed by Panasonic and they are utilizing NJIT as a beta site. The cameras record people entering and exiting the library. The system matches their faces against a database of pre-existing facial images of those who “are banned from the NJIT campus for a number of reasons.” When two images match, the system sends an alarm and public safety officers respond. While the software seems sophisticated–it essentially creates a faceprint “down to skin color, age and gender”–false matches have occurred as described by the security director in the campus news story.
Obviously this type of surveillance at a library can have an impact on someone’s intellectual freedom. If a false match occurred as one entered the library in pursuit of information, the person would presumably have to contend with the public safety officer to correct the false match before getting to the intended book or information resource. It would have a delayed effect on one’s intellectual freedom, but it wouldn’t curtail it, unless a library user objected to the general concept of having someone else create a faceprint of him or her, which seems like the fundamental issue at hand.
It would feel strange if a library started taking fingerprints of patrons who entered and exited just for the purpose of matching them against a state or federal database containing fingerprints of criminals. It feels extreme because one would potentially have to choose between his or her intellectual freedom and his or her desire to control who creates a faceprint. Most people might let the faceprint creation happen just for the sake of ease and freedom of movement. But I don’t think people should be subjected to this kind of decision-making. Do the conditions of eleven property thefts really warrant the heightened surveillance system?
It’s not even so much on a level of expectation of privacy because I think most people understand there’s not an expectation of privacy while in public. It’s more about having a choice in who gets to create and store a face print, which seems like a highly personal and situational matter. Apple and Facebook have face recognition software, but people have a level of say in how they use the software or how images of themselves get distributed. In my opinion, one shouldn’t have to choose between going to a library and having a faceprint created.
Moreover, there’s an “absence of evidence” regarding the effectiveness of security cameras to deter criminality as detailed in Aaron Doyle’s Eyes Everywhere: The Global Growth of Camera Surveillance (pg. 3). There have been no property thefts at Van Houten Library since the system was installed, which is great news, but was it truly because of cameras installed seven months ago? According to Panasonic, they run silently so it’s not as if a potential criminal would know he was being surveilled in that location. I wonder how many alarms got triggered based on accurate matches the system made. I’d also want to know if there’s a process for removing an image from the public safety database used to match faces. The presence of cameras operates on, what Eyes Everywhere describes as “the myth of the rational offender.” Deterrence doesn’t work, Aaron Doyle explains, because criminological research has shown that “in practice offenders do not often approach the decision to offend in a rational calculation” (pg. 7).
I’d also be curious to know how the NJIT community would respond if asked whether or not the presence of face recognition cameras would deter them from using the library in general. If the majority said ‘yes,’ then perhaps the surveillance tactic needs to get revisited with the priority centered on intellectual freedom. If the majority said ‘no,’ then it seems appropriate to proceed, but with effectiveness and good policies in mind.
Allyson Mower, MA, MLIS is Head of Scholarly Communication & Copyright at the University of Utah Marriott Library. She’s very curious about curiosity, what drives people to uncover information, and how libraries of all types create demand for knowledge. As a tenured faculty member, she researches the history of academic freedom — a kind of intellectual freedom — and the history of authorship and scholarly communication at the institution. She provides the U of U community and the general public with information, tools, and services related to both copyright and publishing. Allyson was a Library Journal Mover & Shaker in 2008, was nominated as a 2012 Society for Scholarly Publishing Emerging Leader, and served as the U of U Academic Senate President in 2014. Find her on Twitter @allysonmower.
One thought on “Facial Recognition Surveillance at the Library?”
Interesting post, Thanks for sharing.