Implications of Facial Recognition Technology

Chloe Nelson
8 min readMar 4, 2021

--

Photo by Alexandre Debiève on Unsplash

Not too long ago, my twin sister and I did an experiment to see which of our devices were able to tell the difference between us. Being identical, we often are not able to be told apart from one another by those who do not know us very well. While we feel that we do not look exactly alike, we agree that we do look very similar to each other. We tried both my Windows laptop and her iPhone 12 to see if we could crack into each other’s devices, and we quickly found out that we could easily use our faces to unlock both. It was relatively easy to trick the facial recognition software on the phone and computer, which got me curious about how accurate facial recognition is. This got me thinking about the potential danger that any inaccuracy in facial recognition technology could have if it identified the wrong person in a picture or video.

To start off, facial recognition technology, as defined by the Electronic Frontier Foundation, “is a method of identifying or verifying the identity of an individual using their face. Face recognition systems can be used to identify people in photos, video, or in real-time”. It is essentially any device or software that can be used to identify people based on their facial structure. Many believe that facial recognition is a highly accurate medium, however, that seems to be far from the truth. In a study done by Joy Buolamwini and Timnit Gebru, it was found that in the three different facial recognition technologies they used, there was a huge disparity between the rates of misidentification between Black and White people. The study found that while there was only a .8% error rate for light-skinned males, that rate of error sharply increased when looking at both Black men and women. The error rate for Black women was consistently the highest across all three programs used in the study, with an error rate as high as 36%. Many police precincts use this technology to identify potential criminal offenders and place people at certain crime scenes. In a 2003 article that discussed the idea of implementing facial recognition technology into the police force stated that one positive was indeed surveillance to look for crime suspects or suspected terrorists. This is especially harrowing as there has been a deep rooted history of government agencies using surveillance against people of color based on racist ideologies. The idea of facial recognition and both the positive and negative consequences it holds leaves many people having a stake in the issue of whether it should be used as a trusted technology.

In this case of facial recognition, there are many stakeholders at play. Citizens, both in support and opposition, have a say in what value it actually holds. For some, it may make them feel safer as there are criminals who may be getting caught with this technology; for others it may act as a reminder of institutionalized racism and bias that is so perverse in other parts of life. For those in support of facial recognition technology, they may advocate for its increased use to monitor public spaces. Those in opposition may advocate for a complete stop of its use or a redoing of the systems behind the technology in order to make it more accurate for all people, not just for white people. The technology companies that create the software hold a huge financial stake in this topic, as they profit from any sales of their technology, as well as from any partnering with business and government agencies. Any information that they put out to the public may be done so in a way that highlights all the positive aspects of facial recognition, while glossing over any negatives or inconsistencies with their technology. Lastly, government agencies, such as Immigration and Customs Enforcement (ICE) and Police forces, hold a deep stake in this issue. These agencies may see them as holding value for their investigations and prevention of crime, as they will be able to identify who was in a photo or video taken during the crime. There have already been cases of ICE gaining access to driver license photos in order to find any undocumented citizens they are searching for. Each of these stakeholder’s views and opinions can be looked at through 5 different lenses that help to explain and define them.

Photo by Chris Ried on Unsplash

The first lens that can be used to look at each stakeholder’s belief is the Rights Perspective. This can help explain how citizens, companies, and government agencies approach this topic by examining how it impacts individual rights and the role that it generally places on the people who use it. For citizens who are in support of facial recognition, they may feel that they are creating a safer environment for every citizen in the country, as they may be better protected against crime. However, the dignity of others may be lost through this belief, as it means various agencies and businesses can monitor those without their knowledge. The citizens who are against the use of facial recognition, or feel that its use should be limited, are basing this off the idea that this can be used as a way to invade people’s protected privacy and may even make them feel unsafe as this could negatively affect one group of people over another with misidentification. For businesses that create the software, they are seeing those that are buying the facial recognition technology as consumers. There is a very transactional balance for this viewpoint, as they are more of a focus on selling a product, rather than thinking about the negative implications that it may hold. For government agencies, their view may be a way to protect more citizens and increase their overall well-being and safety. However, there is a side of this that is favoring one group’s well-being over another, as it can be negative for those who are often misidentified with the technology.

When looking through each view through the Justice/Fairness perspective, we are evaluating how an increase or decrease in the use of facial technology may be treating individuals differently or may be treating one group better than another. With any citizen, business or government agency that advocates for its continued or increased use, there is a large issue on how this technology treats everyone that is being evaluated with it. There is an unbalance in the system, and until that unbalance is solved, it will only act to reinforce and worsen any bias that is held in surveillance of citizens. For those who support it decreased use, it may help contribute to a drop in misidentification of those suspected to be in a crime scene, as a different technology would have to be relied on in order to identify suspects,

The Utilitarian perspective has us looking at each viewpoint by examining how facial recognition works, what outcomes are supposed to be positive or negative, and whether any harm caused by the technology may be outweighed by its positives. For those who are in support of facial recognition, they do feel that any increase of public security may outweigh the percentage of people that are misidentified with the technology. Also, the outcome of facial recognition is inherently to implicate someone in a crime, as it is used to place a person at a crime scene. For those that are opposed to facial recognition, there is the belief that the percentage of people that are misidentified is too high and negates the positives that it may hold in identification. The decreased use of facial recognition technology may be justified as it means that other forms of identification, such as fingerprints or DNA, will need to be relied on.

With the Common-Good perspective, the beliefs and viewpoints are evaluated on whether it benefits all groups of people or just some and if it’s creating a positive outcome for anyone that uses it. The advocates of increased use of facial recognition may find that there is a benefit for all people. For government agencies, they may feel that they are helping to keep dangerous people off the street and are preventing undocumented citizens from living in the United States. For those in opposition of facial recognition, there is the perspective that facial recognition is inherently biased and uneven in terms of those who actually can benefit from the technology.

Lastly, the Virtue perspective evaluates whether something is leading us to a better society, what character traits inherently come from choosing this position on facial recognition. For those with the belief that facial recognition is an overall positive technology, they may feel that the technology is creating a deterrent of crime that will lead to an overall safer community. For those in opposition, they may view the technology as something that does not create a better society, but instead reinforces already seen institutionalized racial disparities.

When looking at either argument of supporting or advocating against the use of facial recognition, the two perspectives that I feel best suit this are the Rights perspective and the Utilitarian perspective. Much of this debate involves the idea of surveying people without their knowledge or expressed consent. Many of the pictures used to identify those with facial recognition are driver’s license photos. This means that those who have a license will be more likely to be placed into the facial recognition system than those who do not have their license. This puts certain groups of people more at risk to be misidentified than others. This is not protecting rights as citizens, as there is a large emphasis on creating a role for citizens that revolve around always being a suspect for crime. Also, with the Utilitarian perspective in mind, we are seeing that the goal of facial recognition technology is that someone is identified correctly. However, when there is a high percentage of misidentification with Black people compared to the much lower rate among White people, there is something inherently wrong with the technology being used. With all this information in mind, I feel that it is best to decrease the use of facial recognition technology, and it cannot be used until the percentage of misidentification is virtually zero.

There are many reasons that people choose to support or oppose facial recognition technology. For some, it can feel like a safety cushion that can be used to deter crime and create a safer community. For others, it showcases that there is an increasing presence of police and government agencies within our communities that lead to a worsening of institutionalized racial bias. Whether you support facial recognition technology or not, you must be aware of the alarming rate of misidentification of Black people that occur from these machines. The consequences held for individuals that are misidentified for a crime they did not commit hold much larger consequences than the ones found by two twins being able to unlock each other’s phones and computers.

Sources:

Bedoya, A. (2016, January 18). What the Fbi’s surveillance of Martin Luther King tells us about the Modern Spy Era. Retrieved March 03, 2021, from https://slate.com/technology/2016/01/what-the-fbis-surveillance-of-martin-luther-king-says-about-modern-spying.html

Buolamwini, J & Gebru, T. (2018). Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification. Retrieved March 03, 2021, from http://proceedings.mlr.press/v81/buolamwini18a/buolamwini18a.pdf

Chappell, B. (2019, July 08). Ice uses facial recognition to sift state driver’s license records, researchers say. Retrieved March 03, 2021, from https://www.npr.org/2019/07/08/739491857/ice-uses-facial-recognition-to-sift-state-drivers-license-records-researchers-sa

Face recognition. (2021, February 15). Retrieved March 03, 2021, from https://www.eff.org/pages/face-recognition

Woodward, J. D., Virginia, & Rand Corporation. (2003). Biometrics : A Look at Facial Recognition. RAND Corporation. Retrieved March 03, 2021, from http://web.b.ebscohost.com.ezp1.lib.umn.edu/ehost/detail/detail?vid=0&sid=a558cda8-6ac1-49e8-b12b-7b130163e3b5%40sessionmgr102&bdata=JkF1dGhUeXBlPWlwLHVpZCZzaXRlPWVob3N0LWxpdmU%3d#AN=81628&db=nlebk

--

--