Changing the Face of Identification

Facial Recognition Technology is prevalent across the globe. Here’s what you need to know.

Over the last twenty years, pop culture narratives centred around facial recognition technology, such as Person of Interest, Minority Report, or Black Mirror, have captured the imaginations of a global audience. A world in which the use of facial recognition technology is common might sound like a science fiction screenplay, but the reality is closer than most realise. In fact, many people already interact with facial recognition technology on a daily basis tagging friends in photos on social media, unlocking phones, and speeding through airport security. These examples of simple, consensual, and convenient interactions with facial recognition illustrate some of the ways life can be enriched by technology. However, facial recognition technology is not as perfect as what is portrayed in movies and television. What happens when technology misidentifies a target or over-identifies people when they expect personal privacy?  What happens when powerful governments or corporations use people’s face data to surveil and amass data?

 

Facial Recognition: What is it?

Facial recognition technology gives software the ability to scan an image or live video to identify a person’s face and match it with another image or video of the same person. An algorithm, or a step-by-step procedure for calculations, “learns” about faces from thousands of individual images. When a user uploads a photo, the computer examines the personal unique attributes and compares them to the other images in its “bank”.

How does it work?

Facial recognition software identifies these individual facial features and compares them to others within the database, usually providing a percentage match statistic of confidence.

  • Step 1: A face picture is uploaded to the system.
  • Step 2: Biometrics enable a facial recognition system to map facial features from a photograph or video. Like a human fingerprint or a retina, every human face is unique. There are many distinguishable characteristics that facial recognition software can use to identify an individual. Some systems can identify up to 68 characteristics! Facial recognition technology can distinguish facial features such as the distance between the eyes, the width of the nose, depth of the eye sockets, shape of the cheekbones, and even the length of the jawline.
  • Step 3: The system compiles distinguishing facial features into a mathematical formula, known as a “facial signature”. This signature is compared to a database of known faces. How many Australians have images of their faces in databases? On October 5th, 2017 the National Facial Biometric Matching capability was established. This agreement allows agencies in all jurisdictions to access passport, visa, citizenship, and driver’s license images. Although access to the database is limited to entities such as police and security agencies, it does illustrate how involved the average citizen is in the process of facial recognition.
  • Step 4: The A.I. issues a report on facial matches and percentage match. Previous generations of facial recognition software relied only on 2D imagery. Variations in the photograph such as angle, light, or facial expression greatly affected accuracy. This yielded a high rate of failure. The new generation of facial recognition software can utilise real-time 3D images to identify the subject. Since most photos are not taken under controlled conditions, utilising a 3D model that is not affected by lighting or angles is much more accurate.  Incorporating special sensors, it can even be used in darkness!

 

Why You Need to Know About This Technology

When discussing facial recognition software, many issues arise:

  • Security/Privacy: even without knowing or agreeing to it, your facial data can be collected and stored. Hackers can also access and steal that data.
  • Scale: as facial recognition technology spreads, so will your facial signature. How will you know who has access to it?
  • Ownership: the only face you own is the physical one. Digital images are more complicated. Social media networks can hide rights to ownership clauses within user agreements.
  • Safety: what if someone takes a picture and uses it to discover your identity?
  • Accuracy: a case of mistaken identity could be a disaster. What if you end up being a suspect due to a partial facial match or an erroneous identification?
  • Rights: how can you preserve your privacy if governments or corporations track you? The days of “off the grid” or anonymity are over.

Who’s Using Facial Recognition Technology?

Facial recognition has a variety of applications and is already being applied in daily life.

Travel: International airports use facial recognition technology to secure immigration and some airlines even allow with facial recognition software.

Social Media: Cell phone cameras and social media sites identify faces in photographs and suggest tags in uploaded photos. For example, Apple’s (AAPL) Face ID feature on its latest iPhones utilises facial recognition technology to identify the user’s face. Facebook, Instagram, Twitter, and Snapchat all operate their own forms of facial recognition software.

Security: Businesses can scan customers to identify known shoplifters or prevent fraud, such as retail outlets, banks, and casinos, but also to track the purchasing behaviours of “good” customers to learn more about consumer preferences.

Law enforcement agencies: With access to thousands of images from various agencies of law enforcement, police have plenty of data to sift through to locate criminal suspects and missing persons.

Governments: Across the globe, governments are adopting facial recognition technology. The United States, UK, Australia, China, Germany, India have all adopted facial recognition software for law enforcement, border control, and other uses. Although most of these programs have garnered some resistance, China has received the most noteworthy news coverage based on its extensive deployment of facial recognition technology.

Since 2015, the Chinese government has been building the world’s most powerful facial recognition system. The system currently operates on a local level, such as police, city, or provincial databases, however, the potential is overwhelming. The stated goal of the project is to identify each one of China’s 1.3 billion citizens within three seconds. The project received additional attention when information surfaced that the government was in the process of implementing a “social credit” system by 2020, which will rely heavily on facial recognition software. In addition, the use of facial recognition software to monitor the minority Uyghur Muslim group has also raised global concern about the reach of these technologies.

China has set its sights to become a global leader in artificial intelligence by 2030. Huawei recently purchased the IP rights to facial recognition systems designed by a renowned Russian tech security developer, making it clear that China has no plans to slow down its AI race.

Individuals: A lone tech enthusiast used facial recognition to cross-reference women’s photos on social media with faces on porn sites. He claimed to have created this program to help men determine whether or not their girlfriends acted in adult films. Although the public has generally identified this as a privacy nightmare, without regulation on the matter, individuals will continue to apply facial recognition software to similar datasets for bizarre and possibly disturbing purposes.

The Need to Protect Communities

Although facial recognition systems are more accurate than ever, there are still many examples of inaccuracy and bias. The South Wales, UK police posted data from a pilot of facial recognition software in surveillance at the Champions League final game in Cardiff last year. They logged 173 true face matches and wrongly identified 2,297 people as suspicious. That’s a 92 per cent false positive rate. Assistant Chief Constable Richard Lewis explained why the South Wales force is the first to use this technology in the field:

“The world we live in is changing, and with that comes a need to change the way we police. We are investing in ensuring our officers have the tools and technology needed to most effectively protect our communities. As technology evolves into the future, so too will the way our police force operates.”

Victims of alleged facial recognition failure, such as Ousmane Bah in New York, would most likely disagree. Ousmane, an 18-year-old university student was arrested for thefts he didn’t commit. He asserts that Apple’s use of facial recognition for store security is responsible for him being arrested due to his name being mistakenly linked with the face of the real thief. NYPD officers came to Bah’s home at 4 AM, only to find out that he looked nothing like the man on the arrest warrant. Apple maintains that they do not use facial recognition technology in the company’s stores. Although the case remains open, Ousmane is hardly the first person displeased with facial recognition software.

From the facial recognition fail at the launch of the iPhone X, the racist photo identification errors of Google andNikon, to the American Civil Liberties Union’s test of Amazon’s Rekognition software, which identified 28 members of US Congress to criminal mugshots, there are many incidents that indicate facial recognition technology is not ready for widespread implementation by law enforcement.

To many, such as JetBlue passenger MacKenzie Fegan, a perfectly accurate form of facial recognition software also may not be any easier to swallow. After encountering Jet Blue’s facial recognition check-in, she took to Twitter to share her concerns:

“I just boarded an international @JetBlue flight. Instead of scanning my boarding pass or handing over my passport, I looked into a camera before being allowed down the jet bridge. Did facial recognition replace boarding passes, unbeknownst to me? Did I consent to this?”

 

Although JetBlue’s response was quick and transparent, the incident did bring up potential questions about the technology and its use. It’s nice to shave a few minutes off a run through airport security, but facial recognition software is capable of much more than verifying a person’s face matches the one on the ID card. The massive amount of data available could allow users of facial recognition technology to know your name, address, contact details, waist circumference, preferred Starbucks drink, and dating history within seconds. Imagine how much more valuable data mining could be if you were no longer anonymous? Such concerns about power and privacy have highlighted the need to understand this technology more fully.

What You Can Do

At the moment, options are quite limited. Pending the invention of facial recognition blocking technology, it is now a part of daily life. Some social media platforms, like Facebook, allow users to opt-out of facial recognition settings. Limiting how many pictures are posted on social media can also lower the chances of identity theft. As regulation develops, the voice of citizens will be vital in shaping policy surrounding this technology. In May, San Francisco’s Board of Supervisors approved a ban on the use of facial recognition tech by city agencies, including the police department. The Stop Secret Surveillance Ordinance is the first ban of its kind in a major American city and the seventh major surveillance oversight policy for a municipality in California.

 


 

Technology in itself is neutral.  How it is used determines whether technology helps or harms. While law enforcement and government security agencies view facial recognition software as a tool for good, victims of misidentification or non-consensual surveillance see it as a harmful overstep. Facial recognition technology is an amazing invention and could be used for a variety of beneficial causes. However, the debate surrounding civil rights, law enforcement use, and bias raise several concerns that affect society as a whole, especially for the more vulnerable members of communities. Although the technology is still at its early stages, it is important to consider these factors when deciding how to treat technologies with such vast potential as facial recognition software.

You may also enjoy


Subscribe to
the Prototype