A CCTV camera caught you and you have no clue about it. Next thing you know, authorities are accusing you of a crime you didn’t commit because you resemble the perpetrator. No, this isn’t the screenplay of an action movie. Sadly, it’s today’s reality.
If you think it can’t happen to you because you live in a small town and no CCTV cameras are watching over your every move, think again! Facial recognition tools have managed to invade many business areas worldwide. Unless you’ve never uploaded any personal pictures, odds are facial recognition tools already have your face in their database.
Let’s take a closer look inside how facial recognition software works and what you can do to evade becoming a target. You’ll learn:
- How facial recognition works
- How businesses use facial recognition tools
- How facial recognition is flawed
- How you can protect your privacy and occasionally bypass facial recognition
What is Facial Recognition?
Facial recognition is a biometric identification tool that uses machine learning to identify people in photos, videos, or in real-time. The software takes images from a camera or other digital devices and converts them into data. Then, it cross-checks them against pre-existing facial data on a large database.
The most advanced facial recognition tools can achieve 99.97% accuracy when the images are in ideal condition. Yet, accuracy will fall to a 90% match rate or less with poor lighting and positioning. Law enforcement and security officers initially used the software, but now other business areas like airports, border crossings, and sports stadiums started applying facial recognition.
In some cases, though, facial recognition crosses normal boundaries and goes beyond its initial scope. Here are a few examples.
Human Rights Violations
In China, authorities use facial recognition to identify and track members of the Uyghur ethnic minority. It seems Chinese authorities detain the Uyghurs in reeducation camps, among other abuses that infringe on their human rights.
Civil Rights Violations
In the US, Immigration and Customs Enforcement in collaboration with the FBI use a digital dragnet technology. The dragnet searches for suspects among millions of faces in State driver’s license databases, sometimes without a court order – which the law requires for all searches involving these databases.
Privacy Rights Violations
An investigation led by the Financial Times uncovered how researchers at Microsoft and Stanford University have secretly collected and publicly shared huge data sets of facial imagery without subjects’ knowledge or consent. A program called Brainwash used people coming to a San Francisco café as subjects to develop algorithms of facial recognition tools. Researchers have taken down data sets of Brainwash, but whoever had previously downloaded them can still exploit and play with chunks of photos.
Why have these institutions reverted to using this software? Out of all biometric assessments, facial recognition is the most intuitive. After all, the human eye easily recognizes a face, versus a thumbprint.
Let’s see exactly how facial recognition software works.
The Stages of the Facial Recognition Process
Facial recognition tools use computer algorithms to pick out your specific, distinctive facial details to create a face template. The difference between a photograph and a template is that the latter includes only specific details to distinguish one face from another. Facial recognition systems use different techniques to identify people in challenging conditions like poor lighting, low-quality image resolution, and high angles.
Facial recognition software works like this:
1. Face Detection
Surveillance cameras identify and locate the image of a face, alone or in a crowd. The system captures the image and can determine if you’re looking straight ahead or sideways.
2. Face Analysis
The software analyzes and inspects your face’s geometry. It measures key details: the distance between your eyes, the distance from your forehead to your chin, the shape of your cheekbones, and your lips, ears, and chin contour. Most facial recognition technologies rely on 2D images, and they can create matches with photos in the public domain or from a database.
3. Converting the Image to Data
The system transforms your image into a set of digital data and then into a “faceprint”. This numerical code is unique – just like fingerprints.
4. Finding a Match
The software now compares your faceprint with faces from a large database. Once it finds a match, the process is complete. Some facial recognition systems even calculate a probability match score between you and specific face templates stored in the database.
You can easily imagine how facial recognition tools work by comparing them with Facebook-tagged photos. On Facebook, any picture tagged with a person’s name becomes a part of Facebook’s database. Facebook also can use these photos for facial recognition.
9 Reasons Some Use Facial Recognition
Law enforcement agencies mainly used facial recognition to find crime suspects. Still, since the software proved increased precision, several companies and business sectors have started to use it.
Here are some of the most common facial recognition uses today:
1. Find Missing Persons
Each time law enforcement adds new pictures of missing people in the database, the facial recognition system alerts the police when it recognizes them in an airport, retail store, or other public spaces.
2. Reduce Retail Crime
Many stores use facial recognition to identify known shoplifters, organized retail criminals, or people with a history of fraud. The system matches photographs against databases of criminals and notifies retail security personnel when potential suspects enter the store.
3. Improve the Shopping Experience
Facial recognition technology claims to bring several advantages to your retail experience. For example, it could recognize you and suggest products based on your purchase history. “Face pay” technology also allows you to skip long checkout lines and enjoy a quicker payment process.
4. Make Faster Payments
Biometric online banking is a futuristic use of facial recognition. That way, you could authorize transactions by simply looking at your smartphone or computer. That’s one less password to remember!
MasterCard may have signaled what was to come in 2016 with its Identity Check Mobile. Cardholders take a selfie on their smartphone and quickly authenticate their identity within the retailer’s card network.
5. Enhance Patient Care
Healthcare providers have been testing this technology to help access patient records, streamline registration, detect emotion and pain, and even help identify genetic diseases.
AiCure is an example of an app that makes sure patients take their medication according to their doctor’s prescription. The system uses a tablet or smartphone’s camera combined with an AI motion sensor system that analyzes the data. If patients don’t take the medication correctly, the app sends real-time alerts to patients’ doctors.
Experts believe that when biometric technology becomes less expensive, facial recognition use within the healthcare sector might increase.
6. Track Students or Workers’ Attendance
Some schools in China use facial recognition to check students’ attendance in class. Tablets scan students’ faces and match them to photos in a database to check their identities.
In the US, several schools have employed facial recognition to ensure students’ safety. Facial recognition tools match security camera footage to a database that includes sex offenders, unauthorized staff, suspended students, and other threatening people. Still, some US states introduced laws to ban facial recognition in schools because they could turn into surveillance of students and could potentially be used for disciplinary purposes.
Some companies also use this technology to track employees’ attendance when they sign in and out of their workplace.
7. Recognize Drivers
Some car companies have been testing facial recognition to replace car keys. You’d simply unlock your car through facial recognition software, and the system would also remember your seat and mirror position preferences and radio station presets.
8. Monitor Gambling Problems
Gambling companies use facial recognition to keep a clean reputation and protect their customers.
For example, Australian casinos are using facial biometrics to deny entry to problem gamblers into their locations.
9. ID Passengers
The US Customs and Border Protection applies biometric technology using physiological details to ID flight passengers in several airports. Also, biometric facial scanning, which is optional in many cases, enables flight passengers to use their faces as their tickets. It may gain popularity as more people find ticket-free travel advantageous.
These examples technically portray the bright sides of facial recognition that make our lives easier. Problems come when these tools fall in the wrong hands, or they’re misused.
Concerns Around Facial Recognition Technology
Facial recognition software has what it takes to provide social protection and security, among many applications, as it catches lawbreakers. That said, it also inevitably captures innocent people’s images in various situations or places, which can be intrusive, invading their right to privacy.
Here are some main facial recognition tools’ drawbacks.
They Could Turn into a Surveillance Method
Some worry that facial recognition, surveillance cameras, artificial intelligence, and data analytics create an ideal setting for mass surveillance. These threaten your individual freedom.
That’s why companies like IBM, Microsoft, and Amazon have quit using facial recognition.
They Make Mistakes
Despite being highly accurate, facial recognition data isn’t error-free. Based on recent research, facial recognition works best with white males and isn’t as accurate with women and people of color.
Aging can also increase error rates. An old photo of someone’s face in the database will stay the same while the actual person ages. It makes it difficult for facial recognition to distinguish the two. Changes in the camera angle or appearance, such as a new hairstyle, could also lead to errors.
They Infringe on People’s Right to Privacy
Ethics and privacy are the most controversial matters. Often, facial recognition systems store people’s pictures without their consent.
In 2021, the European Data Protection Board proposed a ban on using facial recognition technology in public spaces to preserve people’s freedom. They also drafted a comprehensive legal framework for applying AI tools. The European Parliament backed up the proposal, but the ban only becomes official once lawmakers in the European Union approve it.
They Lead to Massive Data Mining
Facial recognition tools use machine learning technology. This requires testing and “learning” how to use massive data sets. What better place to find a large photo database than the online world?
For example, anyone who wanted to test algorithms for facial recognition tools scraped photos available on Flickr. That was technically legal as Flickr’s image hosting service covered copyright licenses that allowed liberal reuse. In other words, someone could’ve used, or is using your photos to develop facial recognition tech, and you don’t even know it.
They Accuse the Wrong People
Police officers have wrongly accused people of various misdemeanors resulting from facial recognition failures. That’s what happened with Robert Williams who got arrested because of an improper match from facial recognition software. The police incriminated him for stealing expensive watches from a luxury retail store in Detroit in October 2018. The security footage couldn’t accurately identify the thief’s face as the suspect was wearing a baseball cap. The Detroit police used this footage within a facial recognition software, and it matched the image to Williams’ driver’s license photo.
After police interrogation, they understood that Robert Williams wasn’t the man they were looking for.
Facial Recognition and The Pandemic Challenge
It’s fair to assume that facial recognition has a lower accuracy when most people worldwide wear a facial mask indoors and outdoors. I hate to break it to you, but this isn’t the case.
Many companies have seen it as a challenge to develop even better technologies. Researchers have been using online photos of people wearing masks to build image databases intended to improve facial recognition algorithms.
In Japan, NEC had been working on a facial recognition system to detect the identity of people who wear masks because of allergies, way before the pandemic hit. The system focuses on the face’s uncovered areas, such as the eyes and forehead, and achieves a 99.9% accuracy.
Sadly, the Covid-19 pandemic became the perfect excuse to enforce facial recognition on a larger scale, while the surveillance power of this technology could potentially grow.
How to Evade Facial Recognition as Much as Possible
Facial recognition is handy for those who use it. Still, it could be hindering and unpleasant for the targets of this surveillance tool.
While the most advanced facial recognition tools are highly rigorous, you can still dodge them using 5 simple tactics.
1. Say ‘No’ to Facial Scans
If you’re a US citizen, you can opt out of the facial recognition scan if you don’t want an airport to add your photo to the system. You can tell an airline employee at the time of the facial recognition scan, and they’ll manually scan your documents.
You should know that it may involve a few tricks in between, like spotting the surveillance devices in advance and looking for indications that you can skip biometric identification. For the moment, non-US citizens don’t get to choose to opt out.
2. Use Innovative Photo Concealing Apps
Researchers from the University of Chicago have created Fawkes – a software tool for unauthorized use of personal images. Fawkes ‘cloaks’ photos with hidden changes, and tricks deep learning computer models that power facial recognition without making noticeable changes visible to the human eye.
You can simply run all your photos through this software every time you upload or share a photo online. Anyone or any tool trying to scrape these photos will get a highly twisted version of how you look.
3. Turn Off Facial Recognition from Your Device
Apple and other smartphone manufacturers use facial recognition as an identification tool. If you truly want to avoid becoming the target of any facial recognition tool, don’t use this feature. After all, it’s not mandatory, and you can always resort to using a passcode to unlock your phone.
4. Don’t Tag Photos on Social Media
Facebook and Google created their own facial recognition database, thanks to users (maybe even you) who voluntarily tag their photos on various apps. If you don’t want these companies to monitor and use your face in their databases, then don’t tag photos. If others have added your photos, be sure to un-tag them.
5. Use a VPN
A VPN allows you to browse anonymously and protects your digital identity. Since it removes tracking, it’s incredibly difficult for anyone to figure out who you are. That reduces the information linked to your photos in facial recognition databases.
The problem with facial recognition technology is the fact that as we speak, we’re missing some clear boundaries. While it’s clear that facial recognition tools sometimes make serious mistakes that endanger your privacy rights, they’re still used on a large scale. To make this worse, many countries lack clear laws and regulations regarding the applications of this technology.
You can’t and shouldn’t always try to hide your face. Still, keep in mind that at least in the online world, you can limit your posts and don’t attach personal photos as much as possible.
Whenever you suspect any public or private entity uses your photos or video footage without your consent, kindly ask for clarifications and request to opt-out.
When was facial recognition first discovered?
In 1964, a US research team presumably started to use computers to recognize the human face for a project funded by an intelligence agency. Using around 800 physical images (two photo sets, each of 400 pictures), they programmed a computer to memorize and compare different versions of people’s faces. The photo sets consisted of people’s faces of different ages and varying in head rotation.
They concluded a machine could recognize a face better than humans.
Is facial recognition legal?
The US doesn’t have a specific law that regulates facial technology use, even when law enforcement and police officers use it. Some US states have proposed several bills to ban facial technology.
In the EU, facial recognition and biometric data technically fall under the General Data Protection Regulation (GDPR).
China just recently (August 2021) passed a law on the application of facial recognition, leaving the Chinese citizens the option to opt out of this technology where possible.
Overall, countries don’t have clear regulations set in place about who, when, where, and how anyone should use this technology properly and under legal terms. That’s why so far, public and private organizations use facial recognition as they see fit.
What’s the difference between facial detection and facial recognition?
Facial detection is a broader term that encompasses facial recognition. Facial detection means a system can identify if a human face is present in an image or a video and can apply it to facial recognition. Facial recognition does more than simply identify a human face; it compares a face to images from a database and looks for a match, trying to determine a person’s identity.
How do I know if someone used my pictures for facial recognition training?
If you’ve uploaded or shared any personal photos recently, researchers may have used them to build facial recognition systems. An online tool called Exposing.ai can help you figure out if anyone scraped your pictures.
Does facial recognition technology make you anxious? Do you think it’s a good idea for these tools to be put to use in public spaces?
Let me know in the comments section below.