Facial recognition necessitates privacy protection in public

By DENG XIANPING / 06-03-2020 / (Chinese Social Sciences Today)

In recent years, some schools have installed AI-powered facial recognition systems in classrooms in an effort to facilitate the management of daily attendance and classroom discipline, but the move has caused discomfort among many students. Photo: FILE


In recent years, some schools have installed AI-powered facial recognition systems in classrooms in an effort to facilitate the management of daily attendance and classroom discipline. However, the move has caused discomfort among many. Some students complained that their privacy has been infringed upon, while the schools maintained that students shouldn’t raise privacy concerns because the measure was aimed at urging them to study harder. Related departments held that classrooms are public places, so the accusation of privacy violation is untenable.


Binary status of privacy broken

In ancient times, forms of technology were assembled, usually consisting of objects that already existed in nature or that were processed by rough means. In technical operation, humans were regarded as part of the assembly. Life was similar to work, and the operation interface of life was not significantly different from that of work. Humans were simply tools, with no or little privacy. Even if they had privacy, it was basically to conceal potential disgrace.

Privacy was a reflection of free will. More privacy indicated stronger free will. When it came to realizing certain goals, the stronger free will the “tools” had, the more unpredictable the realization of the goals would be.

Since modern times, technology has been discovery-based. In other words, humans discover inherent laws of nature to strengthen their capacity of technical operation. Technological development has segregated people through artificialities. With the expansion of artificialities, humans have become more powerful in conquering and exploiting nature, while the segregation of people has been aggravated.

In today’s world, even next-door neighbors are not familiar with each other, a situation that differs fundamentally from the society of acquaintances in the past. This is a result of technological development. Scientific rationality has continuously challenged humans’ mythological cognition of the world, with human values highly advocated. Modern privacy has thus become conceptualized.
At the same time, education and training are increasingly vital to human technical operation.

Moreover, technical operation aims to finish work goals. The more complex a technical operation is, the higher requirement for operation is imposed upon humans, and the more oppressive technology becomes to humanity. Their life is totally different from work. The individual requirement for privacy and the compliance with goals through technical operation have brought into being a dichotomized pattern of privacy comprising life and work.

Daniel Solove, a professor of law at George Washington University, pointed out that privacy is usually viewed as a binary status, as the society is divided into two totally different domains: public and private. Being in a public place suggests no expectation of privacy. “As one court observed, appearing in public ‘necessarily involves doffing the cloak of privacy which the law protects,’” Solove said.

Generally, due to the dichotomy between private and public domains, traditional privacy studies tend to exclude privacy issues from public domains.

The binary approach also manifests in the dichotomy between public personal information (PPI) and non-public personal information (NPI). Herman Tavani, a professor emeritus of philosophy at Rivier University, drew a distinction between PPI and NPI. PPI mainly refers to individual work information, which is personal but also presented in public places. And NPI means information about life that appears in private domains. 

However, the development of information technology has started to break the binary status of privacy. It informatizes not only artificialities, but also the operation interface of artificialities and displays them explicitly.

Traditional technology attends to technical goals, while information technology begins to value the process of the realization of the goals. The purpose of installing the facial recognition system in classrooms is to monitor the process of study in order to enhance learning efficiency. It makes classroom technology more intelligent, yet it directly peers into facial information and interprets it.
As classrooms are informatized, surveillance will pay more attention to the representation of public personal information. The information is about both study and personal life, leading to the continuity between public and non-public information. Consequently, information that is traditionally considered private can now be examined as public personal information, hence causing privacy leaks.


Traditional privacy redefined

In ancient society, public domains had no exact boundaries. They were acquired by occupancy. Who occupied more resources owned more discourse power. From behavioral perspectives, public domains were culture-based. The combination of humanity and technology gave rise to behavioral norms for different fields. Renowned German philosopher Jürgen Habermas defined public domains in ancient times as a representative public sphere that belonged to a few. For the majority, there was no privacy.

With the intervention of modern concepts and lifestyles, public domains began to transform. The public sphere is an umbrella name for all public buildings, places and facilities used by the public for work, study, business, culture, social interaction, entertainment, sports, visits, healthcare, hygiene, rest and travel and to meet part of their living needs. It serves certain purposes with definite temporal and spatial boundaries. The orientation of realizing a certain goal in a certain time and space is disconnected to individual life. It is because of the disconnectedness that many experts admit no privacy in public domains.

In the information age, the installment of facial recognition systems in public places such as classrooms has substantially extended the scope of the public sphere. The original temporal and spatial limits have been broken by information. Traditionally in public places, micro facial expressions and movements have been beyond capture by human eyes, not to mention analysis.
When facial recognition is able to supervise students’ micro facial expressions and analyze them, the definition of classrooms as a public place has changed. These micro facial expressions are concerned with non-public personal information. The traditional scope of individual privacy is now incorporated into the public sphere.


Therefore, installing the facial recognition system in classrooms cannot be justified by the statement that “classrooms are public places, so there is no issue of privacy infringement.” When the system can monitor human micro facial expressions and movements and tries to regulate them, it has penetrated the private life of individuals. Modern technology is entirely capable of inspecting private life, just as modern archaeology has restored the lifestyle of the ancients by means of carbon-14, paleopathology, DNA detection, mass spectrometry and other micro technologies.


Privacy protection in public

Technology pushes the autonomous development of humanity, the public goal to prompt students to focus more attention on their lessons is reasonable, and privacy protection is necessary, so coordinating the relationships between the three is key to promoting the application of new technologies.

A normal approach is to restrict both supervisors and the supervised to limit invasion of privacy. However, it is obviously impractical to expect the supervised to fully regulate their previously private behaviors under the surveillance of a facial recognition system. Nor is it feasible to surveil supervisors and limit their use of the facial recognition technology to a public purpose only. The public behaviors of supervisors can be monitored effectively, but their private deeds cannot, otherwise their privacy will be violated as well.

To separate the system from its operation interface might be an effective approach. The capturing of human micro facial expressions can be regarded as a contact effect between the system and humans and realized by the operation interface. Supervisors identify micro facial expressions by operating the system.

It is advisable to isolate the operation interface used by supervisors from real facial recognition. The identification and interpretation of micro facial expressions is realized by the technology, and supervisors are responsible only for seeing the final result.

When supervisors need more information involving privacy, they can obtain it with the consent of the supervised and after the decoding of technicians. Technicians will be forbidden from attaining firsthand decoding information. In short, the aim of privacy protection is achieved by extending the technology itself.

Can the application of facial recognition technology facilitate human learning? If properly applied, it surely can. The proper application means students can leverage the technical data to analyze their own performance and improve their learning methods and attitudes accordingly. But pure surveillance is likely to inhibit their freedom of and enthusiasm for learning and hinder their studies.

Deng Xianping is a research fellow from the Institute of Philosophy at Guangdong Academy of Social Sciences.

edited by CHEN MIRONG