Facial recognition technology (FRT) is a biometric technology that—if the New York City Police Department (NYPD) implements it in the New York City subway system—could have significant benefits of preventing violent crime, deterring terrorism, helping investigate past crimes, locating missing persons, providing assistance to individuals with special needs, and integrating with other technology platforms to allow for greater efficiencies in policing. Since the NYPD successfully uses license plate reader (LPR) technology to increase public safety with an existing legal and policy framework governing its use, and because FRT is similar to LPR technology—in terms of law enforcement use of the technology as well as benefits and challenges—this thesis presents a comparative analysis of the two technologies.
From this analysis, the following questions can be answered: To what extent are LPRs and FRT analogous, and how can the NYPD’s use of LPRs provide a roadmap for public support of real-time FRT? Currently, the NYPD does not use automatic or real-time FRT, but specially trained detectives assigned to the Facial Identification Section use FRT to investigate past crimes and assist detectives in the field. Other law enforcement agencies outside the United States, however, are using or evaluating the effectiveness of real-time FRT, which has significantly improved in recent years, in part because of the development of high-definition video, advancements in storage capabilities, and the ability to evaluate faces in real time. Modern FRT systems can also recognize an individual with varying facial expressions. Therefore, a person can be accurately identified in a facial recognition database, even if the facial expression is different from the original image contained in the database.
Communities benefit from LPR technology, despite privacy concerns and the contention of critics, such as the American Civil Liberties Union, which assert that law enforcement agencies should not be able to collect and store information on law-abiding citizens who are not suspected of criminal wrongdoing. This thesis proposes that although privacy considerations may exist, the benefits of LPR technology outweigh these concerns. LPRs and FRT are similar types of technologies that perform similar functions. They both scan the image of an unknown variable and attempt to match it against a known variable contained in a database, collecting all information—without bias—on license plates or persons. Furthermore, both technologies use “hot lists,” which compare the scanned image against information or images contained in local, state, and federal databases. Computer software then compares the scanned image against collected data or images. Finally, the information, such as images of a stolen vehicle or wanted person, received from both technologies requires human verification. Like LPR technology, FRT—especially when used in real time—has significant benefits and can act as a force multiplier for limited NYPD resources in crowded environments such as the New York City subway system.
The NYPD has developed sound policies with LPRs and has already mitigated the challenges to minimize potential harm due to misuse and violation of civil liberties. In addition, LPRs are generally acknowledged as a common and publicly accepted law enforcement technology, largely in part because the NYPD proactively addressed many of the risks. The NYPD addressed LPR data collection, retention, and sharing through robust and clear policies. In addition, the police department addressed potential misuse of LPRs to eliminate their ambiguities and clearly define acceptable practices. Through comparative analysis, this thesis determines that real-time FRT could help law enforcement deter terrorism, prevent violent crime, identify wanted individuals, find missing persons, and assist in mental health situations and post-event investigations. This thesis also addresses the litany of challenges in the use of FRT—privacy concerns as well as false positives, false negatives, intentional circumvention of real-time FRT, and law enforcement misappropriations—and identifies the concerns over how law enforcement collects, shares, and disseminates personal information obtained from facial recognition software. This thesis concludes that real-time FRT can help the NYPD meet its mission by reducing fear, increasing resiliency, and adding a layer of protection for citizens riding in the New York City subway system.
 Ben Eisler, “ACLU Concerned Automatic License Plate Readers May Invade Privacy,” WJLA News, July 30, 2012, http://wjla.com/news/local/aclu-concerned-red-light-cameras-may-invade-privacy-78301.