Author: Annie Wang
From: Irvine, CA, USA
Should the Use of AI & Facial Recognition Be Restricted?
As the use of artificial intelligence continues to expand and integrate into daily lives, a social dilemma arises: the ethics behind the use of facial recognition due to the potential invasion of racial equality, privacy, and security. The secret use of facial recognition by the police and other agencies has advanced to a serious form, mass surveillance. Yet facial recognition technologies are currently not in the matured phase as they misidentify women and minorities at a much higher rate. In addition, mass surveillance using racial recognition without consent from individuals undermines the basic principles of liberty, privacy, and equal protection under the 1st, 4th, and 14th Amendments. As a result, it is crucial to initiate policies and legislations to protect these civil rights. These policies should, first, ban generalized face surveillance that reveals personal information by arbitrating scanning citizens without their notice or consent, and second, ensure training datas are inclusive in all races, gender, and age groups in the future development of facial recognition and artificial intelligence.
Freedom and Privacy
The current use of facial recognition in a form of mass surveillance is invading individual’s rights to freedom and privacy. As depicted in the novel 1984 by George Orwell, the current use of face surveillance in society may turn the society into a dystopia if there are no regulations drafted. The first Amendment grants individuals civil rights and freedom of speech while the fourth Amendment protects the rights to privacy, yet face surveillance scans the face of individuals and enters them into datasets without the individuals’ consent or notice. The government is monitoring citizens in forms like real time video without their knowledge in virtually unrestricted ways. A legislation should be crafted to ban general surveillance that could reveal personal information without individuals’ consent. No parties should be allowed to use facial recognition in secret. In addition, individuals should not be forced to submit their face data and their face should not be entered into datasets without their consent. Legislations should authorize the use of facial recognition for investigation purposes on only those who are suspicious, but should not be targeted at the general public. There is a huge difference between using facial recognition versus mass surveillance. These policies will help to draw the line clear between these two forms. It is crucial for policies to be taken now because unregulated facial recognition is simply too powerful, as it undermines the basic principles of freedom, privacy, and security.
Racial Disparities
In addition to the issue of mass surveillance, the potential racial disparity brought by the use of facial recognition also makes it necessary to initiate policies to ensure that the training data in the development phase of facial technology encompasses all races and to not use facial recognition as “the” reason for arrests and accusations. Found by various research and studies, “facial recognition technology misidentity women and minorities at a much higher rate than white males”. This issue will increase the risk of racial and gender bias and guide the society in the wrong direction. The fourteenth amendment provides equal protection under the law, but as a result of racial and skin type biases in artificial intelligence systems, several cases of false arrest and accusation happened due to misidentification. As a result, actions need to be taken now. A policy that requires data of all skin types to be included for any future development of facial recognition or artificial intelligence is needed. In addition to skin types, equal representation in training data for women, minorities, and youth is necessary to ensure the biases within AI is minimized. Other than policies on the development of AI technology, policies are needed on the use of facial recognition. Facial recognition technologies should only be allowed to serve as a source of evidence, but not as a reason to arrest and accuse a citizen. A person should not be arrested immediately just because he or she is detected by facial recognition. Judges and police will need to make judgements based on other evidence presented, as well as a personal review of the suspect. These policies on the police’s use of facial recognition must be enforced before police can use facial recognition as a tool, because the capacity for abuse and technical immaturity poses risk for misaccusation and misidentification, especially for marginalized communities.
Conclusion
Although facial recognition or AI technologies may be useful to maintain the security of society by identifying criminals, the risks behind unregulated facial technology is too great to not initiate policies. There needs to be policies to ban the use of unregulated facial recognition, especially in mass surveillance without individuals’ consent or acknowledgement, and individuals should not be forced to enter their face into the datasets. Furthermore, there needs to be equal representation of all races, genders, and age groups in the development of facial recognition. As we move forward in the path of technological developments, ethical concerns like privacy and racial inequalities need to be addressed at equal rates as breakthroughs are made.
Author: Annie Wang
Annie is a rising high school senior who is passionate about exploring the intersection of math, computer science, and economics. In her free time, she enjoys playing music, doing art, and spending time with her family.
Ivanova, I. (2020, June 12). Why face-recognition technology has a bias problem. Retrieved August 15, 2020, from https://www.cbsnews.com/news/facial-recognition-systems-racism-protests-police-bias/
Committee, O. (Director). (2019). Facial Recognition Technology (Part 1): Its Impact on our Civil Rights and Liberties [Video file]. Retrieved August 15, 2020, from https://www.youtube.com/watch?time_continue=2683&v=jLmzEFkbNsg&feature=emb_logo
Wong, Q. (n.d.). Why facial recognition's racial bias problem is so hard to crack. Retrieved August 15, 2020, from https://www.cnet.com/news/why-facial-recognitions-racial-bias-problem-is-so-hard-to-crack/
Great reminder that even though artificial intelligence innovations are super cool, we need to keep in mind the question of whether they're ethical or not.
The racial disparity part is interesting. It just shows how our society makes white males the "norm," and everyone else as "others" (this is not meant to be an attack on white males). Hopefully that will change one day!