The History and Controvery of Facial Recognition in China and America
Facial recognition technology is a system that is able to recognize and confirm someone’s identity through a video source. A basic overview of how facial recognition works is that it analyzes features of someone’s face and crosschecks it with the system’s database, which usually contains some form of ID. Facial Recognition, along with technology of all kinds, is continuously growing and advancing with time. Although it once started off as a computer application, facial recognition has grown to be used in mobile applications and through other methods. Both China and the United States have implemented forms of facial recognition, although they are quite different from one another. The main difference has to do with the privacy and security aspects of this technology. Many people are unaware of what is being done with this data. Because of this, there is a lot of controversy and unsolved mysteries regarding usage and the sociotechnical aspects of this technology.
So, how exactly does facial recognition technology work? People are able to recognize who someone is based on their features such as the width between their eyes, the size of their lips, and expressions they make. Facial recognition does the same thing, except using an algorithm. Although different technologies have different methods to their systems, there are some basic processes. First, your face is recognized and captured from a video or a picture. The image capture may show your face in any angle. Because humans and computers are not able to recognize faces the same way, engineers have come up with ways to get over this hurdle. Phones, such as the iPhone X and XS, that use facial recognition, have the user move their face around to know what the user looks like from different angles. There are also other systems that take 2D images of your face and put it on a 3D model to achieve the same result. In China recently, a new rule went into effect on December 1st, 2019. Mobile phone users in China that install a new SIM into their phones must also submit a facial recognition scan. This rule mandates phone companies to implement this technology to know exactly who is registering the SIM card.
Second, your facial signature is determined through your facial features that are crucial to differentiating your face from other faces. Third, your facial signature is taken and cross checked with other facial signatures. Approximately 117 million Americans have their faces within police databases. Lastly, the facial recognition system determines your identity.
To give some background information about facial recognition in China, it has the largest video surveillance network in the entire world, and in the next two years, they want to expand it to 600 million cameras. Currently, China uses facial recognition in many different ways. One of these ways is through tracking if citizens are breaking the law by jay-walking. Near intersections, there are some stations which can detect who is jay-walking. The algorithm cross checks with the person’s ID, and publicly shames them by putting their face on the screen for all to see. China uses it through many other means of citizens living their daily lives.
The history of facial recognition is one that grew very fast within a short period of time. In the 1960s, Woodrow Wilson Bledsoe created a system called the RAND tablet, where faces could be recognized by hand. A RAND tablet was used to place coordinates on different aspects of the face such as mouth, nose, hairline, and eyes. This information was then put into a database, where they could select a person that best resembled that profile. However, this system was very limited, but gave hope that facial recognition was possible. Then, during the 1970s, AJ Goldstein, LD Harmon, and AB Lesk used 21 subject markers to identify faces. Compared to the RAND tablet, this new system could be done automatically.
In 1988, Sirovich and Kirby incorporated linear algebra into facial recognition and created a method called Eigenface. They presented that feature analysis on facial images could be compiled to create basic features. The Eigenface concept was later expanded in 1991, when it became another example of automatic facial recognition.
However, not all of the growth of facial recognition was good. In the 2002 Super Bowl, facial recognition was utilized by authorities on large crowds to detect potential criminals. They received a lot of false positives, which in turn resulted in backlash.
In 2010, Facebook began using facial recognition to autotag people in photos. In 2014, they declared that they could recognize faces just as well as the human eye, or 97.6%. Facebook still does this to this day and has advanced a lot since then.
In 2017, the iPhone X came out and facial recognition was advertised as a feature to make the user’s iPhone more secure. It was a very popular phone, which just proved that facial recognition was something that many were growing accustomed to in this time period.
SenseTime is one of China’s artificial intelligence companies. Founded in China, it has offices in Japan, Singapore, and in the United States. This company develops algorithms for employees to check in during events and meetings. It can also rank the beauty of a person, as well as determine the age, sex, and mood of them. SenseTime created a system for a “Smart and Safe City”. This system tracks citizens by labeling their clothing, sex, and age.
This is not to say that there are not any immoral or negative things being done with personal data in China. An article on The Guardian called China’s hi-tech war on its Muslim minority tells the story of a Muslim, Uighur man named Alim, who was forced into a detention center and punished for things he was completely uncertain about. In 2017, when Alim landed in China from studying abroad, he was taken by the police to complete a “health check”. He was then taken to a detention center, where he stayed in a cell with 20 other men. In his two weeks there, he was deprived of food and sleep, and went through hours of verbal abuse and interrogation. Getting put into a detention center could be simply because you travelled outside of China, have a relative abroad, or posting about Islamic practices on social media apps. Even having connections to people doing these things could get a person in trouble. There is an obvious bias within this system. Even though Alim was released from the detention center, his data followed him everywhere. The solution police officials gave to him was to, “Just stay at home.”
Comparing China’s facial recognition to America’s, there is a big difference. It does not compare to the horrifying things happening in China regarding mass surveillance, but there is still surveillance and facial recognition being used in the United States in ways citizens may not exactly approve of. The FBI has a facial recognition database of 641 million images of Americans, which can be searched easily without any warrant.
This past summer, FaceApp, a mobile application that can transform a user’s face to make it look older or younger, smile, and change gender, was repopularized. A rumour began to go around that the app may have been using the pictures taken to store in a Russian facial recognition database. However, this was proved to be false. This rumor brought up many other concerns about apps and their usage with user data. Facebook uses DeepFace, a deep learning facial recognition system that collects data from pictures of users, unless they have opted out of the software. Over many years, users have uploaded different pictures of themselves in many different stages of life and location. Because of this, Facebook does not require any extra set of photos to complete its data collection. In the future, Facebook could potentially sell this data and facial recognition services to airlines and credit card companies.
Compared to China, a lot of the means of facial recognition in the United States is right in our hands — our mobile phones. Americans are obsessed with using their phones and companies know this, so they could easily implement some sort of facial recognition feature on an app that people are completely unaware of. The terms and conditions of an app are not usually read by many, and it could be one way that people overlook what data they are giving the app access to when they install it.
As this technology continues to expand, there should be laws and regulations from local, national, and international levels. Facial recognition has come a long way and made enough advancements where big mistakes like during the 2002 Super Bowl do not occur in the United States. There should be local laws because every city has different circumstances and it may not relate to every other city and state. As for international, there also needs to be laws put in place so people’s lives are not put at danger because of unknown and invalid reasons.
Words Cited
Facial Recognition, www.doc.ic.ac.uk/~hh4017/History.
Symanovich, Steve. “How Does Facial Recognition Work?” Official Site, us.norton.com/internetsecurity-iot-how-facial-recognition-software-works.html.
