Edward Claughton, CEO
PRI Management Group
While in the gym this morning, I overheard a man and a woman discussing last night’s presidential debate. One of them said, “I just think she’s easier on the eyes; how the person looks, I think, is just really important you know”. The other one agreed. They continued to chat about the “optics” of it all, personal appearance, and quite frankly, about nothing that matters in the selection of a President. It is amazing how far we have strayed from assessing the issues, the facts, the things being said, and from questioning them. Like a sponge absorbing everything it’s given, critical thinking, it seems, has gone out the window.
In the same vein, when it comes to today’s issues in criminal justice, decisions are all too often made without much forethought and in response to ill-informed perception. Such is the case today for a revolutionary technology which can literally have profound positive results in the reduction of crime. Facial recognition.
Facial recognition really isn’t new. Body-worn camera companies included video redaction functionality in their systems, including a feature which would track faces and aid the user in blurring the image. This is facial recognition.
That same technology has since advanced to work better in other systems, and for use in catching criminals. But, the knee-jerk reactions began, as did the sensationalized headlines, political whim and thus a few cities putting the brakes on the technology because 1) they don’t understand how it works and 2) they don’t really understand the nuances of probable cause and criminal investigations. There is important context to consider.
First understand that police video, surveillance video seized as evidence from a robbery, mugshots, driver’s licenses, etc. are all records. They fall under an agency’s retention schedules (at least they should). Managing all of this administratively speaking requires sound records management operations including the fact that once a video has lost its legal, historical, administrative and fiscal value, it should be destroyed.
Why is this especially important with facial recognition? These systems are used to identify suspects and if someone is determined not to be the suspect, the video/photo of the innocent person in question should be managed accordingly.
Valid studies have shown facial recognition technology, which by the way is a type of technology that is “smart”, meaning it learns the more it is used, inherently has a higher error-rate when it comes to correctly matching photos/video of people with darker skin. Likely due to natural human demographics, the fact there are more light-skinned people than dark, the technology isn’t as smart with minorities. This however is soon to be corrected as the algorithms advance and the technology is fed a more and more diverse population. It is nonetheless a valid concern. However, when measured without due consideration to traditional policing methodologies, bad decisions are made. Case in point…
Recently in Detroit, according to The Detroit News, “Police officials have revised the proposed policy governing the use of facial recognition software, removing one of its most contentious provisions: The ability to use the technology to scan faces in real time if there’s a terror threat.”
Remember the Boston Marathon bombings? How about September 11, 2001? Would you want your local police force scanning the suspects at the marathon or the airport to catch the offenders?
San Francisco, a once beautiful, now decrepit city overrun with open-air drug use, homelessness and unsanitary conditions, continues down a path of political destruction. The City voted to ban facial recognition; a knee-jerk reaction.
Facial recognition is here to stay. This artificial intelligence technology provides law enforcement the ability to rapidly identify suspects whose identity is otherwise unknown. Once someone has been identified as a possible match, the investigation proceeds to the next step- developing confirmation of who the person is, that they were in fact the person involved in the crime via traditional investigative techniques, and establishment of probable cause for arrest.
Police have since the beginning of photography, forensics sketches, and the advent of security cameras compared suspects to mugshots, driver’s license photos and open-source/online photos, all by hand. Such overwhelmingly time-intensive work prevents the identification of most suspects captured on camera. Now, it can be done in minutes, if only our elected officials would take the time to make informed and responsible decisions.
Recognize the controls law enforcement has adopted to ensure the responsible use of this technology through policy, training, auditing, multi-layered review and confirmation of identification by third parties. Understand that it is merely one step of the investigation- one however that could immediately identify someone who has committed a heinous crime, and may be about to commit another.
PRI has developed a comprehensive artificial intelligence technology integration service which allows agencies to adopt facial recognition programs responsibly through customized policy, training, stakeholder education, systems integration and support, using Veritone Identify and Veritone Redact. Contact us to get up and running with these systems.