Facial Recognition Technology
August 1, 2022

In a 4-2 vote last week, members of the New Orleans City Council voted to allow the New Orleans Police Department to employ facial recognition technology while investigating certain crimes. No doubt in response to the largest crime wave since Hurricane Katrina, the result reverses a 2020 city ordinance prohibiting the use of this controversial technology. So why is this important? 

Alongside private companies, governmental entities have a long, contentious past of mining for and collecting civilian data without consent. Justification often presents itself as being for the benefit of national security, public safety, and even economic stability. Facial recognition technology (FRT) relies on this type of data and is developed using an agglomeration of millions of images, like DMV photos and posts made on social media platforms, which are collected and dumped into a database. Biometrical facial analysis of these images can then be scanned using images captured on surveillance software, like CCTV, in an effort to find a “match.”  

Law enforcement agencies, such as the N.O.P.D., assert that this type of technology has the potential to assist in conducting faster investigations as well as stop and or prevent future criminal activity. While there may be some truth to this logic, for a software used in the name of public safety, its unreliability is wildly problematic. 

FRT is subject to differential errors and depends on a variety of factors. Some of these errors are based on things like gender and race, while others are a product of poor image quality. FRT operating settings can play a large part too. Confidence thresholds, for example, are functions based on pre-programmed metrics determined by a law enforcement agency that specify how close of a match that agency requires the system to search for.

The lower the setting, the greater the likelihood that an identification will be inaccurate.  The unreliability and variable use of this technology, creates concerns regarding the potential for misidentification, wrongful arrests, as well as human rights and privacy violations. There’s even the potential for hackers to gain access to the technology and use the information for nefarious purposes.  

Without stronger privacy protections, what can be done about it? 

Not many people are keen on the idea of their faces being recorded and stored for unknown use, or the notion that they’re always being watched. A quick walk around any New Orleanian neighborhood will bring you face to face with corner cameras. Due to the differences between jurisdiction on federal, state, and local levels, however, regulation may be tricky.  One option would be for Congress to enact federal privacy laws that would regulate use of FRT on a national level. Under such a scenario, law enforcement agencies would be given contextual guidelines and measures for when use of facial recognition software is appropriate and how it can be utilized, which could include a warrant requirement.