What’s in a Face?
Weighing the risks and benefits of facial recognition technology

Facial recognition technology (FRT) has grown greatly in capability and range of uses in recent years. It has provided solutions in medicine, consumer applications and in security and law enforcement. However, the increased use of FRT has given rise to concerns about it misuse along with concerns surrounding privacy, civil liberties and bias. Several states have enacted proposed laws to regulate the use of FRT, but the laws are not consistent from state to state.
In order to take advantage of the power behind FRT while safe guarding privacy and civil liberties, there must be a federal law that protects the entire country against its risks. In order to better understand what constitutes feasible federal regulation, a more detailed examination of FRT is required.
The Racist History of Facial Recognition
Facial recognition traces its roots to physiognomy, the practice of using a person’s outer appearance to draw conclusions about character. Cesare Lombroso, a 19th-century Italian scientist and surgeon, recorded precise measurements and physical traits to conclude that criminals were born criminals. His proposed theory was an attempt to justify racist beliefs long held by northern Italians who felt that southern Italians were racially inferior.
In the 20th century, German biologists like Ernst Haeckel used the same type of rationalization of racial hierarchy to vilify foreign groups. This same type of theory of physiognomic types was then used by the Nazis to justify the persecution and genocide of Jews. Physiognomy has been long discredited and labeled a pseudoscience. But that hasn’t stopped modern day researchers and startups in their attempt to use face data to assess character and intelligence. Artificial Intelligence and machine learning have not only driven a rebirth to the now debunked science of physiognomy, they’ve also provided a sense of objectivity around its unethical use.
How Facial Recognition Technology Works Today
Facial Recognition technology can be divided into face data and facial recognition. Face data is any information that is related to an identifiable person and includes images from an array of sources, including images provided by the persons themselves, government identification records, social media, web sites, criminal mug shots and surveillance. Facial recognition describes the process of extracting data, picking out faces from the data, identifying unique features of the face and then comparing the face to find a match. Machine learning’s use of convolutional neural networks (CNNs) has made FCR accurate, automated and fast.
Benefits of Facial Recognition Technology
Companies and organizations have developed consumer-facing applications of facial recognition Technology that have improved the customer experience. FRT applications include enhancements in safety and security, access and authentication, tremendous leaps in diagnosing diseases, and targeted marketing and customer service.
Of the listed applications of FRT, those related to safety and security are often the most hyped benefits associated with the technology. Law enforcement agencies use FRT for finding missing persons, identifying wanted criminal and for early threat detection. In medicine, FRT can be used to diagnose diseases that cause detectable changes in appearance and is expected to become an invaluable diagnostic tool for all types of conditions.
On a consumer level, FRT can be used as a tool for locking personal devices and accessing accounts or services, like those in banking. Face recognition requires no hackable passwords, and the technology can be used at physical branches and ATMs. FRT also has valuable applications in retail. Using your face to pay can drastically cut down check out lines. Customer experiences can also be improved through customized shopping experiences.
Drawbacks of Facial Recognition Technology
While the benefits of FRT are tangible, let’s look at a couple pitfalls that come with the technology.
Informed consent: Facial recognition has drawn the attention of lawmakers and civil and privacy right advocates due to its expanded use. Facial recognition technology is typically deployed without consent, by both the private and public sector. As a result, the collection and subsequent use of the face data raises civil and privacy right concerns. Unlike other forms of account authentication, face data is unique and non-replaceable. Therefore, security concerns with regards to facial recognition become paramount. Once face data has been hacked, a person’s privacy and security are jeopardized as their biometric data can no longer be used as a secure identifying feature.
Bias: Facial recognition technology should be expected to show accuracy for specific uses to demonstrate that overall benefits overcome risks. However, this is not the case. CNNs are trained with face data that contains mostly white male face images. As a result of this lack of diversity, facial recognition algorithms don’t perform well when asked to identify non-white faces. The ACLU demonstrated this bias when they tested Amazon’s facial recognition tool, "Rekogntion", by comparing members of Congress’ faces to a data base with images of people arrested for a crime. The software erroneously matched 28 faces to the data base, with the false matches being disproportionately people of color.
Facial Recognition Technology: Ethics and Privacy
Ethics: Organizations that use FCR technology use a variety of moral excuses to rationalize the unethical use of facial recognition technology that can perpetuate biases and hinder privacy and civil liberties.
Privacy: Beyond identification, FRT can be used for surveillance in such a way that that doesn’t require the knowledge, consent or participation of those being monitored. Since surveillance cameras are now ubiquitous, people have become accustomed to being recorded in public. These cameras, which belong to private companies, law enforcement and government agencies are no longer noticed and are thought of as just being used for security purposes.
Individuals don’t fully grasp how FRT removes the anonymity that is associated with being recorded in public spaces and allows organizations to not only identify individuals, but to also analyze their facial data and draw insights from it, without consent. This aspect of FRT has drawn the attention of lawmakers and privacy right advocates due to its expanded use by both public and private sectors. But what are States and the Federal Government doing about the expanded use of FRT?
Current State Laws Surrounding Facial Recognition: The WPA 2020 Model
States have been grappling with how to regulate organizations that are developing and implementing facial recognition technology. Current laws targeting facial recognition include Illinois’ Biometric Information Privacy Act (BIPA), California’s Consumer Privacy Act (CCPA), and the Washington Privacy Act (SSB 6281), which will soon be voted on by the State Senate.
Of all the current state laws, the protections afforded by the Washington Privacy Act go beyond those in the CPPA, the only all-encompassing consumer privacy law in the United States. To better highlight the law’s strengths, it’s informative to compare the Washington Privacy Act, the CPPA and the EU’s General Data Protection Regulation (GDPR).

Permissible Uses of FCR
The Washington Privacy Act provides a good template for Congress to use in its effort to regulate Facial Recognition Technology. It addresses protections for commercial uses of FRT while building on the GDPR’s model of protecting individual rights and ensuring that company obligations are met. Since FRT laws differ from state to state, having Congress pass a law at the federal level will ensure that all Americans are protected from the misuse of FRT. Privacy and civil liberties can be further protected by the passing of the Facial Recognition Technology Warrant Act, which would require a probable cause warrant to track a person’s public movements for over 72 hours, and limits tracking to 30 days.
Any federal law must provide a strong ethical framework to help businesses navigate grey areas. This framework should emphasize fairness and inclusivity, accountability, trust and adaptability. With the right regulation and protections, facial recognition technology can make the world safer, more convenient and smarter. The wide range of permissible benefits that FRT can bring to society includes diagnosing diseases, aiding in law enforcement, facilitating secure transactions and improving customer experiences. Having the right safeguards in place will help ensure everyone safely and securely benefits from FRT’s advances.