Facial Recognition Software is Not as Scary as You Think
It’s amazing how many imagined technologies from old science fiction books and movies are becoming realities these days. Recently, one of these “future-world” technologies exploded into tangible present-day prominence: facial recognition. As intriguing as the technology is, there is definitely controversy centered around facial recognition preventing many from taking advantage of its benefits for crime mitigation. In some cases, hesitation to utilize facial recognition could potentially cost lives. The purpose of this writing is to provide reassurance and guidance to those considering a facial recognition solution, and by doing so provide you with information to help you to move forward with confidence.
In this article, we would like to dispel a few myths surrounding security-based facial recognition software and increase your knowledge in this reliable and safe technology.
Debunking the Biggest Arguments Against Facial Recognition: Reliability & Privacy
Facial recognition detractors’ main areas of concern tend to be (1) a fear that the AI (artificial intelligence) behind facial recognition software is not reliable – and (2) when it IS reliable that it will jeopardize an individual’s right to privacy.
We have to admit those fears would be well-founded if we were talking about the last generation of facial recognition. In the past, the technology used facial geometry to measure parts of your face and it could be reverse engineered to figure out what your face looked like. That would be a problem, for example, if someone was trying to pretend they were you to get into your bank accounts or ingress through your company’s loading dock door to steal equipment. To make matters worse, the initial technology’s accuracy wasn’t as defined, resulting in many false positives. Today’s facial recognition software and algorithms are much more precise, significantly reducing misidentifications.
Today’s facial recognition software no longer uses facial geometry. Instead, the AI assigns a mathematical number that represents your face. That means it’s doubly difficult for humans to hack, and it’s much more accurate. It’s not a human telling the software what a face looks like; it’s an arbitrary number, which protects privacy.
Furthermore, today’s security-based facial recognition technology works without a master database. Each company establishes and creates their own individual database that is only accessed and managed internally. There is no danger of an image of a particular face being indiscriminately circulated (let alone shared in association with a name, birthdate, Social Security number or bank account number).
Prime Communications, Inc. (PCI) partners with a popular facial recognition solution, AnyVision. AnyVision explains that today’s facial recognition systems keep data within the boundaries of one location and doesn’t touch the Internet, so it is much more secure.
“Unless you want to share your data with other locations in your organization, such as stores in a retail chain or schools in a district, the information is not accessible by anyone else. And we don’t share our clients’ data” AnyVision explains. The one exception is the occasional sharing of data with law enforcement organizations in special circumstances. For example, a school might share the identity of a person who is a known danger to others so they can be apprehended quickly.
Here’s the Thing About Privacy: You are Already Identifiable
At Glenelg High School in 2018, a group of young adults wearing ski masks spray-painted racial slurs on the sidewalk around the building and on the stadium. Security cameras were in place, but the masks made facial recognition – by computer OR human – impossible. However, because the students had phones in their pockets that were linked to the college network via username, they were identified and charged with the crime.
Corporate Facial Recognition technology can only gather the image of your face. It doesn’t automatically know your name or any other information about you, unlike the data being harvested from your phone. In addition to being trackable, like the phones of the college students, laptops and phones use facial recognition to unlock them, and every time you use this feature the manufacturer is capturing an image that they have a right to and can do with what they want.
Your image and/or other identifiable information is being captured every day. Facial recognition software for security isn’t collecting or storing any new data, as your phone is. It’s simply applying AI to see if locally captured images are the same as stored images. PCI advises that free phone apps almost always involve someone harvesting some kind of data about you and selling it. Your information is often tied to information identifying you, including credit card numbers if you’ve purchased goods using your phone.
Why do we accept intrusive technology on our phones and computers while being nervous about security cameras capturing our images? The answer is simple. People are willing to risk being identified via the technology in their phones because in their minds the benefits outweigh the risks.
The benefit of a camera capturing our images for security purposes is not as immediately clear to consumers. However, the fact is that cameras and facial recognition software in schools, public buildings and places of business can keep us safe through deterrence, detection and prosecution of criminals who mean us harm. That’s a pretty good ultimate benefit!
For organizations, weighing benefits and risks is the key to determining whether and when to use facial recognition for security. PCI’s VP of National Accounts, Brian Freeman explained that we all have a right to expect some level of privacy. Everyone would agree cameras should not be placed in bathrooms or dressing rooms, for example. Commercial enterprises likely see cameras as a necessary part of their business at entrances and in lobbies and other public areas, and there isn’t an expectation of privacy in those places. “To decide whether to mount cameras in an area of your building,” he said, “ask yourself where an expectation of privacy ends and a need for safety begins.”
Proactive vs. Forensic Facial Recognition
According to AnyVision, there are two sides to security-based recognition software: proactive and forensic. In both uses, a video management system is constantly recording video in the background from a variety of views at any location.
Forensic uses of facial recognition include running that video footage against facial images of known criminals after a crime is committed. This usage seems to be acceptable because the benefit is immediately apparent. Everyone wants to catch a criminal.
In contrast, proactive facial recognition involves using real-time scanning to alert security officers if a known person of interest comes into view. It is this usage of facial recognition that causes some alarm with those who don’t understand that facial recognition is now reliable and safe.
The problem in the minds of some is that they believe facial recognition software constitutes a breach of the Fourth Amendment of the U.S. Constitution – protection against unreasonable search and seizure. In Freeman’s experience when people are discussing facial recognition with him, they regularly mention Minority Report – a movie in which an organization is using technology to predict who will commit a crime and then arresting them based on that prediction.
Technology doesn’t really work like that. In fact, court cases have established that such “searches” are reasonable when public interest is high and the intrusion into people’s lives is small.
Today’s facial recognition software is not that different from wanted posters in the wild west. Only, instead of relying on infallible humans to remember seeing the face of a cattle rustler and report to the sheriff as they did then, we’re using technology to ‘read’ faces and report to modern-day law enforcement, which is much more reliable. The Innocence Project reports mistaken eyewitness identifications contributed to approximately 71% of the more than 360 wrongful convictions in the United States overturned by post-conviction DNA evidence.
Just as security personnel have the right to visually scan a room to search for faces of known criminals, they also have the right to use cameras to help them digitally scan the room.
For an organization, the decision whether or not to use facial recognition, among other considerations, requires balancing need vs. budget. You can put the software on every camera in a stadium, for example, but it’s probably not worth the time and money. Most companies are starting with entrances and chokepoints to catch criminals before they get on the premises.
Facial Recognition Regulations are Not in Place – Yet
With the technology being so new, most jurisdictions have not yet formulated regulations specifically addressing the use of facial recognition software, and as far as we know there is no federal law governing facial recognition at this time. Because the technology is misunderstood it has been banned from three American cities for police use. Illinois is currently the only state to have enacted a Biometric Information Privacy Act. This act requires any company harvesting biometric information (fingerprint, iris or facial geometry) to obtain written consent from all parties involved. Much like GDPR in Europe it also ensures that companies gathering this information house it in a safe way so the data can’t be compromised.
The security industry is in the process of analyzing needs, and producing guidelines and best practices companies and institutions can follow. AnyVision supports the regulation of this technology, but it is encouraged that policymakers focus on the proper use of the technology instead of just banning it outright.
All the Good That Can Be Done with Facial Recognition
For many security professionals, it comes down to the good that can be accomplished using facial recognition. As long as the software isn’t attached to mass databases or used illegally, how can you argue with positive outcomes such as these:
- Working hand-in-hand with law enforcement, schools can identify sex offenders the instant they show up on campus.
- Retailers can identify known criminals instantly and call for assistance, which can help take these people out of the equation for all other retailers, reducing violence and saving many dollars.
- One study found that facial recognition software decreased the time it took for police to find a potential suspect from 30 minutes to three.
- Casinos can watch for known card sharks. In one case, facial recognition helped apprehend a person who had stolen millions of dollars.
- Children lost in a department store can be found fast to minimize the opportunity for successful kidnapping. When a child is missing, every second counts.
- The Boston Marathon bomber may have been found faster with the help of facial recognition. There are many other historical tragedies that might have gone differently if perpetrators could have been identified faster.
- Disgruntled employees can be denied access to workplaces before they attempt violent acts.
There is still a great deal to be sorted out by our society to ensure the public that facial recognition will not be – and cannot be – used for nefarious purposes.
For now, we suggest these best practices:
- Engage your legal team early to make sure your desired use of facial recognition is appropriate and legal, understand the limitations and risks, and determine a protocol for putting selected persons on your watch list.
- Make facial recognition protocols part of your standard operating procedures and carefully think through the reasons that it can be justified as a reasonable business practice.
- Think through all the “what ifs.” Make sure your protocols include thoughtful responses to a variety of situations that might call for the use of facial recognition.
- Choose a facial recognition system that gives you the ability to use existing cameras, offers speedy processing and notification, and is easy to use, including smoothly integrating with other equipment, software and protocols.
How does an integrator like PCI fit in? We can help you leverage existing security solutions, navigate the complexities of behind-the-screen facial recognition software, and use our depth of experience to help you understand the issues and regulations surrounding the use of facial recognition in your industry- particularly if you are in the retail, education or medical industries (our top-tier specialties).
We’ll help you ask the right questions about facial recognition – and get the answers you need. For more information, call 402.289.4126 or email us at firstname.lastname@example.org.