I recently spoke at the Biometrics for Government and Law Enforcement conference in Washington, DC. The more I speak at these events and around the country at our Lunch and Learn events, the more I realize that there are many misconceptions about biometric facial recognition technology. Maybe it comes from too many crime dramas on TV or maybe it comes from reports like the Georgetown Law Study that represented a narrative about bias that the facts simply don’t support. During speaking engagements like this one, I consider it a duty and a privilege to cut through the myths and share the reality of how facial recognition can be an incredibly valuable tool for law enforcement to develop leads and help solve cases.
Separating Biometric Facial Recognition Myths from Reality
While there are many misconceptions to address, I’ll take you through the top four that were discussed at the Biometrics conference.
Fact One: It works.
When I joined Vigilant Solutions after serving over 20 years in the New York Police Department (NYPD), I was asked, “If you had one message that you would like the world to know about facial recognition what would it be?” My response: “I want the world to know that this is a biometric technology tool and not a biometric science.” What do I mean? Facial recognition is not 100% accurate. It is not DNA. But here’s what I do know: After working on countless cases using facial recognition, I am certain it works when properly used in the investigative process to develop leads and help solve cases. At the NYPD, the facial recognition investigation unit I spearheaded has conducted more than 8,500 facial recognition investigations and over 2,700 arrests. I know that without it, many cases would go unsolved. And I know that it has helped bring dangerous criminals to justice, it has saved lives, and it protects officers and communities.
Fact Two: A facial recognition match is NOT the basis for arrest.
This is so important and that’s why we emphasize this over and over: Facial recognition technology provides leads; it is NOT the basis for arrest. The onus always falls on the agency to establish probable cause. Agencies should follow best investigative practices and establish policies for facial recognition investigations including: documenting what you have done, creating an audit trail, and demonstrating a disciplined approach to facial recognition investigations. Critical to that disciplined approach in addition to the background investigation, is a requirement that an analyst individually examine the potential match to determine if the physical attributes match up.
Fact Three: Enhancing images helps to find the bad people.
At the conference, there was a ton of buzz about the ability to enhance images. Some even questioned the practice and asked whether it was a bad to manipulate probe images. The facts show that we are not manipulating but enhancing what is already there to develop a lead. When we use enhancement tools to render a poor-quality image usable for facial recognition investigations, we are simply taking that image and cleaning it up so that it can be used to generate a potential match by the facial recognition software. For example, what happens when you add eyes to a picture of someone with closed eyes and you end up with a potential match that ranks 100? You investigate, you look into the background, and you vet that photo. Take the case of twins. If you put in a photo of one twin the other will show up as a match. Does that mean you arrest the match? No, what you have is a lead to investigate. Isn’t the same thing done when agencies use sketch artists? The find someone who resembles the sketch through physical characteristics but still need to validate the potential match as a viable candidate in the Investigation.
Fact Four: You don’t have to be a graphic design expert to use enhancement tools.
For many agencies, biometric facial recognition is off the table because they falsely believe that you need to either be an expert at graphic design or you have to outsource image editing software and training to use the technology. My time with the NYPD as the Lead Detective for the agency’s first dedicated facial recognition unit taught me that law enforcement needs tools that are easy to use or they won’t use them at all. That experience led to the development of Vigilant’s simple to use image enhancement tools.
We are here to help separate the fact from the fiction, to tell the stories of successful investigations using facial recognition technology and to equip agencies with the best practices and policies for implementing facial recognition. Thank you to everyone that attended my presentation and visited me at the show. Let’s keep the conversation going and share the facts.
Want to know more? Get the Whitepaper: Facial Recognition: Art or Science?