AI and Ethics (2023) 3:369–379

https://doi.org/10.1007/s43681-022-00196-y

REVIEW

“Ethically contentious aspects of artificial intelligence surveillance:

a social science perspective”
Tahereh Saheb1

Received: 20 May 2022 / Accepted: 29 June 2022 / Published online: 19 July 2022
© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2022

Abstract
Artificial intelligence and its societal and ethical implications are complicated and conflictingly interpreted. Surveillance
is one of the most ethically challenging concepts in AI. Within the domain of artificial intelligence, this study conducts a
topic modeling analysis of scientific research on the concept of surveillance. Seven significant scholarly topics that receive
significant attention from the scientific community were discovered throughout our research. These topics demonstrate how
ambiguous the lines between dichotomous forms of surveillance are: public health surveillance versus state surveillance;
transportation surveillance versus national security surveillance; peace surveillance versus military surveillance; disease
surveillance versus surveillance capitalism; urban surveillance versus citizen ubiquitous surveillance; computational sur-
veillance versus fakeness surveillance; and data surveillance versus invasive surveillance. This study adds to the body of
knowledge on AI ethics by focusing on controversial aspects of AI surveillance. In practice, it will serve as a guideline for
policymakers and technology companies to focus more on the intended and unintended consequences of various forms of
AI surveillance in society.

Keywords Artificial intelligence · Ethic · Surveillance · Topic modeling

1 Introduction artificial intelligence (AI) surveillance technologies (Fig. 1).

However, incorporating AI-based surveillance technologies
Artificial intelligence has had a substantial influence has been a game changer in advancing effective measures in
on civilization, whether in the form of algorithms and a variety of industries, including healthcare, transportation,
machine learning models, or robots and autonomous sys- and manufacturing. Others, on the other hand, condemn arti-
tems. Enhancing surveillance and monitoring is one of the ficial intelligence-based surveillance technologies for their
most critical uses of artificial intelligence. At least 75 of the unexpected or intended harmful implications, particularly
world's 176 countries, according to the Global Surveillance in the lives of citizens, and their potential to support anti-
Index (GSI), are actively investing in and deploying artifi- democratic policies and violations of privacy and human
cial intelligence (AI) for surveillance purposes, primarily rights principles [25, 48]. The studies express outrage at
in smart cities, facial recognition, and smart police [17]. governments for violating human rights through the use of
Governments have integrated artificial intelligence into surveillance technologies for pandemic management [48,
cameras, video management software, and mobile phones in 52].
collaboration with technology corporations [35] and normal- As global concerns about the battle between digital
ized biometric surveillance in the control of pandemics, such authoritarianism and liberal democracy sprout [67], aca-
as the current global COVID pandemic [4], Saheb et al., demic researchers from diverse backgrounds address vari-
2021c). According to the GS Index, Chinese and American ous dimensions of artificial intelligence for surveillance; as
technology companies are the world's leading providers of there are conflicting and perplexing perspectives on the con-
sequences of artificial intelligence for surveillance. While
* Tahereh Saheb some studies focus exclusively on the good effects of AI on
[email protected] surveillance, others emphasize the negative aspects of AI
on surveillance.
1
Management Studies Center, Tarbiat Modares University,
Tehran, Iran

13
Vol.:(0123456789)
370 AI and Ethics (2023) 3:369–379

Fig. 1  Countries use American (left) and Chinese (right) Tech companies to supply their AI surveillance technology [17]

As of March 10th, 2022, about 3556 scholarly papers with neglected and overlooked research streams. Theoretically,
the keywords artificial intelligence and surveillance were this study will contribute to the growing body of knowledge
indexed in the Scopus database. As illustrated in Fig. 2, the regarding social science studies of technology and the eth-
number of scholarly research on this subject has increased ics of artificial intelligence. Practically, this study will serve
dramatically since 2017 and peaked in 2021. As illustrated as a guide for policymakers and technology corporations
in the figure, the subject is an interdisciplinary one that has interested in gaining a better understanding of the societal
attracted the interest of scientists from a variety of fields, consequences of surveillance using artificial intelligence
including computer science, engineering, mathematics, technologies and techniques.
medicine, physics and astronomy, and social sciences. This study will attempt to address the following research
This paper, on the other hand, intends to map the scholarly questions:
endeavors of social scientists and humanities working on AI
and surveillance to further non-technical discourse about • What is the social science and humanities perspective on
surveillance and AI and to distinguish conflicting perspec- the knowledge structure of AI surveillance? What scien-
tives on AI for surveillance. The recent upsurge of interest in tific discourse exists around the controversial utilization
ethical AI to emphasize the transparency and accountability of artificial intelligence for unethical surveillance pur-
of artificial intelligence has enhanced social scientists' schol- poses?
arly endeavors [33] to underscore both advantages and disad- • How has social science research on artificial intelligence
vantages of AI surveillance by governments, industries, and and surveillance evolved over time?
corporations. This study intends to map new scholarly efforts • What are the under-researched and marginalized areas
by social sciences and humanities scholars to comprehend within social science studies of AI and surveillance, as
the most frequently addressed issues, thereby discovering well as the potential research strands?

Fig. 2  The evolution of scholarly growth over the interdisciplinary topic of artificial intelligence for surveillance based on papers indexed in Sco-
pus database

13
AI and Ethics (2023) 3:369–379 371

2 Methodology based on modularity, and another to depict the evolution

of keywords using overlay visualization. The items were
On March 9th, 2022, we extracted data from the SCOPUS represented according to their total link strength, or TLS
database. We searched the abstract, title, and keyword for score. The modularity algorithm provides significantly
the terms artificial intelligence AND surveillance. We more strongly related keywords and sheds light on the
limited our study to disciplines in the social sciences and strength of networks [46]. Additionally, the TLS score
humanities and omitted review articles. Only English-lan- considers the number of connections an object has to other
guage articles, conference papers, books, and book chap- objects as well as the strength of those connections [48].
ters were included. We imposed no further restrictions.
This search resulted in the retrieval of 293 documents.
After screening the titles and abstracts of the papers, we 3 Results
determined that 14 were irrelevant. Thus, we did topic
modeling analysis on 279 papers at the end. As a consequence of the co-occurrence analysis of key-
We employed keyword co-occurrence analysis as a words, seven main scholarly subjects were identified, as
bibliometric technique to identify the most influential illustrated in Fig. 3. According to the analysis, the first
scholarly topics. We next conducted a content analysis highly addressed topic is public health surveillance and the
of relevant publications to each topic to complement the privacy of contact tracing apps during pandemics, notably
bibliometric analysis findings. We analyzed the 279 pub- the COVID-19. The second topic discusses video surveil-
lications for content analysis. Co-occurrence analysis is lance and facial recognition technologies, mostly in the
a widely used quantitative technique for determining the transportation industry. Topic 3 refers to studies on military
structure of research and its possible academic relevancy surveillance and artificial intelligence in its physical forms,
[29, 44]. We analyzed and visualized networks using the such as drones, robots, and autonomous vehicles. Topic 4
VOSviewer software. This software is a widely utilized pertains to studies on surveillance capitalism and disease
tool for analyzing and visualizing scientific literature, as surveillance. Subject five is about smart cities, while topic
well as tracing the evolution and knowledge structure of six is about computational surveillance. Additionally, the
scientific topics [62, 64]. We displayed the analysis in two final topic discusses security surveillance. Noteworthy is the
ways: one to depict the themes whose normalization was fact that essentially, the majority of the topics have over-
lapping discussions, but their primary focus is distinct. For

Topic 2: Video Surveillance & Facial

Recognition in Transportation

Topic 1: Public health surveillance &

privacy during pandemics

Topic 7: Security & Data

Surveillance

Topic 3: Military
Surveillance

Topic 6: Computational
Surveillance

Topic 4: Disease Surveillance

Topic 5: Urban Surveillance & Smart Cities

Fig. 3  Co-occurrence analysis of papers yielded in seven influential scholarly topic on AI and surveillance

13
372 AI and Ethics (2023) 3:369–379

instance, while public health surveillance and disease sur- and warfare [32]. Researchers have voiced worry that the
veillance share a number of contentious ethical issues, their automated collecting of personal data has exacerbated the
primary focuses are diverse. As an illustration, the central power imbalance between citizens and governments and
emphasis of disease surveillance is on state and citizen sur- technology companies [41], as well as the secrecy and
veillance, whereas the primary focus of public health sur- lack of transparency of corporations [13]. These concerns
veillance is on privacy concerns. have led to the development of technical solutions such as
privacy by design, privacy by default [26] and differential
privacy [8], and new governance frameworks [7].
4 Topic modeling and content analysis Biometric facial recognition has been identified as a
significant concern, because it creates privacy concerns
4.1 Topic 1: public health surveillance & privacy when employed for citizen surveillance, crime control,
during pandemics activity monitoring, and facial expression evaluation [8],
posing a threat by automating unauthorized access to per-
Human rights violations are facilitated by cutting-edge tech- sonal data such as facial images [57]. A second signifi-
nologies, which empowers governments to collect biometric cant area of worry with AI technology is the use of health
and other personal data and legalize AI-based tracking sys- wearables and applications. In these instances, the gather-
tems in the premise of public health protection [19]. During ing and processing of personal data must be authorized,
pandemics, notably COVID-19, tracking systems have been transparent, and duration and scope limited [26].
employed to conduct epidemiological surveillance of indi-
viduals at risk [65]. However, there is rising concern that AI
surveillance for reasons such as detecting new COVID-19 4.2 Topic 2: video surveillance systems & facial
cases or collecting data from healthy and severely ill indi- recognition in transportation
viduals is being employed for purposes other than public
health management, thereby breaching individuals' privacy The second issue with AI surveillance is the implementa-
[53], decreasing citizens’ trust and voluntary adoption of tion of video surveillance systems (VSM) and facial rec-
these technologies [65]. In some studies, AI-based surveil- ognition in public transportation. The majority of research
lance is described as a form of biopolitics and as a tool for have focused on applications of AI to transportation sur-
enhancing government surveillance and control [60]. Schol- veillance; very few have addressed the ethical and legal
ars have proposed technical solutions to mitigate the ethi- consequences of such applications. Studies identify that
cal ramifications of AI-based surveillance systems, such as employing AI surveillance in transportation is primarily
de-identification and anonymization of data and differential intended to reduce trespassing frequency and deaths [71],
privacy [21, 48], 2021c). Conversely, other scholars believe monitor terrorist activity and suspicious behavior [42],
that a balance should be struck between data privacy and improve public safety and crime-fighting capabilities [14],
public health [34], and that broader privacy laws, such as and develop intelligent traffic surveillance systems [30].
the European Union's General Data Protection Regulation Despite its efficacy, AI surveillance in transportation has
(GDPR), the California Consumer Privacy Act (CCPA), sparked ethical concerns, particularly when the police and
or the US Health Insurance Portability and Accountabil- the government collaborate to execute AI-based surveil-
ity Act of 1996 (HIPAA) Privacy Rule, should be adopted lance systems in public transportation zones [14]. Studies
[26, 53]. More precisely, research indicates that AI surveil- illustrate the intricate interrelationships between human
lance of vulnerable and marginalized populations should and computer authority [56], human rights breaches [1],
be implemented cautiously. Additionally, they advocate and threats to civil liberties and freedoms [14]. Other stud-
for collaborative and multidisciplinary efforts [22]. Due to ies assert that there is no universally applicable human
the fact that governments are developing AI surveillance rights framework or regulatory standards for facial recog-
systems in collaboration with technology companies, prior nition and video surveillance technologies [1]. In light of
research has emphasized the importance of developing gov- smart policing initiatives in public spaces, such as the tran-
ernance frameworks, exercising control over various actors, sit system, human rights breaches are highly exacerbated
and increasing technology companies' social and political when minorities, African-Americans, and children are
accountability [7]. engaged [68]. In some cases, these efforts could lead to the
AI surveillance has been fraught with privacy concerns remedy of unconstitutional practices and racial discrimina-
beyond pandemics. With the emergence of surveillance tion [40]. There have been some studies, suggesting that
societies [50], a slew of worries about privacy and the the ethical concerns surrounding facial recognition tech-
blurring of lines between human–machine interaction have nology should extend beyond privacy and transparency
surfaced [22], as have new forms of government control to include issues of equality, diversity, and inclusion [1].

13
AI and Ethics (2023) 3:369–379 373

4.3 Topic 3: military surveillance politicization of care, anti-democracy and social dissimila-

tory practices, and violations of human rights [48].
The third most significant category of studies focuses on
physical AI, robotics, and AI-based devices in the military, 4.5 Topic 5: urban surveillance and smart cities
as well as military surveillance, as these technologies have
progressed beyond science fiction movies [15]. These tech- The fifth most discussed topic of AI surveillance is smart
nologies may offer tremendous benefits, including increased cities and the collection of personal data for urban plan-
precision and accuracy, as well as automation of military ning. Studies point out anxieties regarding the leakage of
measures and counterterrorism programs; however, some personal information [54], full automation and absence of
researchers have highlighted concerns that autonomous AI- humans[10], and control of city facilities and influence on
based drones and technologies in the military will ramp up citizens lives [70]. As studies have shown, among the most
discriminatory practices and unaccountable military activity sensitive and ethically problematic forms of personal data
on a political and legal level [58]. Furthermore, research in a smart city are customer profiles, time-spatial travel, and
have demonstrated that applications of Unmanned Aerial automated fair collection [9]. Smart city surveillance com-
Vehicles (UAVs) and drones constitute complex technolo- promises human privacy on a variety of levels, including
gies augmented by a variety of technologies [69], neces- identity privacy, query privacy, location privacy, footprint
sitating the development of specialized legal frameworks privacy, and owner privacy [31]. Citizens' privacy concerns
to address civil liberties and privacy concerns [18]. Non- are primarily triggered by their perceptions of city data and
combatant civilians are being killed by military drones, the rationale for its use [63], as well as unlawful access to
igniting much controversy [6]. Previous research has raised confidential data and cyberattacks that disrupt the delivery
concerns regarding armed and autonomous robots' ability to of city services physically. [61]. According to study under-
accurately discern between combatants and non-combatants, taken by [72], other ethical issues confronting smart cit-
or to distinguish between dangerous and nonthreatening con- ies include control and data ownership, friction between
duct [28]. There are a variety of ethical concerns surround- the public and private sectors, social inclusion and citizen
ing autonomous AI-based military technology, including involvement, and subsequent disparities and prejudice.
ambiguity about who is responsible and if these technologies
are capable of unethical or illegal behavior [55]. 4.6 Topic 6: computational surveillance

The sixth topic explores computational surveillance, soft

4.4 Topic 4: disease surveillance computing, and security challenges faced at the time of
the development of computational intelligence. According
The fourth key theme in AI surveillance has been the sur- to [12], computational intelligence can be defined as the
veillance of diseases in the context of capitalism and the design, application, and development of biologically and lin-
emergence of surveillance capitalism. Concerns have pri- guistically motivated computational paradigms, including
marily developed as a result of COVID-19 and government natural language processing, natural language understand-
collaborations with large technological corporations [48]. ing, and natural language generation models. As computa-
Shoshana Zuboff first proposed the concept of surveillance tional intelligence advances, attention is being directed to
capitalism in 2007, contending that Facebook and Google, computational ethics, so that it can be distinguished from
two technological companies, earn financial success and machine ethics and robot ethics [51]. Scholars frequently
help governments with political advertising by selling use soft computing techniques, such as natural language pro-
algorithmic-driven, micro-targeted profiles of individuals cessing (NLP), to promote civility and monitor hate speech,
[26, 66]. COVID-19 provided opportunities for technology cyberbullying, toxic comments, and abusive language [49].
companies to pass legislation that boosted their commercial Furthermore, computational intelligence has been employed
benefit while limiting citizens' freedom of movement and to detect propaganda, fake, and manipulated news [36]. Dur-
access to their personal information [11]. Numerous ethi- ing the COVID-19 outbreak, computational intelligence was
cal questions have been voiced about artificial intelligence's extensively used for disease surveillance [39]. In the wake of
application in healthcare. Scholars dispute whether artificial cyberattacks on computational intelligence applications[23],
intelligence and its applications, such as tracing systems, are security-enabled design techniques and algorithms, as well
desirable for citizen spying or disease surveillance in both as the construction of secured software and the strengthen-
democratic and dictatorial nations [16]. Depersonalization ing of threat modeling during software development, were
and dehumanization, as well as discrimination and disci- put into practice [2, 38].
plinary care, are among these ethical concerns [43]. Fur- Furthermore, computational intelligence has also been
ther ethical risks of AI-based disease surveillance include incorporated into the political sphere for the purpose of

13
374 AI and Ethics (2023) 3:369–379

fabricating propaganda, fake news, and hate speech. In light and unmanned aerial vehicles (UAVs) for object detection.
of this trend, concerns have been raised regarding the vul- Between 2018 and 2020, attention was concentrated on the
nerability of individuals, political parties, institutions, and rise of big data, the Internet of Things, drones, data privacy,
communities, as well as the possibility of manipulating them legislation and regulation, and citizen acceptance of technol-
for destructive reasons [27]. At the dawn of the deepfake era ogy. After 2020, pandemics, COVID-19, contact tracing,
[20], deep learning algorithms and generative adversarial human rights, capitalism, and surveillance in transportation
networks (GANs) aggravated the situation by altering the became popular topics (Fig. 5).
appearance of human subjects on existing photographs and
videos to make them appear like someone else [59].
6 Discussion and literature gaps
4.7 Topic 7: security and data surveillance
We conducted a co-occurrence analysis and a content analy-
The seventh topic pertains to security and data surveil- sis of 279 scientific papers on AI and surveillance authored
lance. In recent studies, concerns regarding the automated by social science and humanities experts. Our investiga-
decision-making capabilities of AI systems have generated tion found seven significant scholarly topics that receive
concerns regarding security and data surveillance among cit- significant attention from the scientific community. These
izens, which has hampered their adoption of these technolo- topics demonstrate the ambiguous boundaries between
gies [37]. A rising tide of datafication and data-driven sur- dichotomous forms of surveillance: public health surveil-
veillance has contributed to a life fraught with uncertainty, lance versus state surveillance; transportation surveillance
with civil society concerned about countering threats posed versus national security surveillance; peace surveillance
by surveillance, data exploitation, and vulnerable systems versus military surveillance; disease surveillance versus
susceptible to cyberattacks [24]. The increasing availability surveillance capitalism; urban surveillance versus citizen
and usage of big data has invaded everyday life, threaten- ubiquitous surveillance; computational surveillance versus
ing citizens' privacy and security in intelligent surroundings fakeness surveillance; and data surveillance versus invasive
packed with technology that extract personal information surveillance.
[3]. Due to the infiltration of big data into every facet of The most distressing topic in AI surveillance is public
human life and the abundance of citizens’ digital footprints, health surveillance and the deployment of COVID tracing
covert monitoring of citizens' behaviors, intentions, and applications, which blur the boundaries between citizen
preferences is now conceivable. According to Forbes maga- and public health surveillance. As the COVID pandemic
zine, the US government secretly ordered Google to provide expanded across the globe in 2019, various countries devel-
information about customers who type in specific search oped mobile-based contact tracing applications to track
phrases [5]; highlighting governments' unlawful access to and halt the virus's transmission. Concerns about privacy
consumers' digital data. and surveillance emerges as a result of these applications'
capacity to autonomously access their users' location and
contacts [45, 47], 2021c), fueling public suspicion that these
5 Historical evolution of surveillance applications were instruments of citizen surveillance. Ethical
concepts considerations are heightened in countries where citizens are
required to adopt the applications. Due to compulsory instal-
This section of the article examines the evolution of surveil- lation, a lack of proper rules, and collaboration with technol-
lance concepts across time. As illustrated in Fig. 4, prior ogy corporations, individuals lacked trust and harbored con-
to 2012, topics, such as vehicle tracking, computer vision, spiracy theories that these applications were employed for
predictive models, border surveillance, and military appli- citizen surveillance and capitalism's empowerment, rather
cations, were among the most linked concepts in scientific than public health surveillance. The significant drawback
study. This means that the experts emphasized the impor- of these studies on artificial intelligence for public health
tance of AI for military and border surveillance. From 2012 surveillance is the absence of studies examining the indirect
to 2014, the development of statistical models and machine effects of contextual factors on the adoption of AI-based
learning techniques gained prominence. Between 2014 and public health surveillance systems. Multiple sociological,
2016, attention was mostly focused on the societal rami- economic, and political considerations, as well as their indi-
fications of artificial intelligence surveillance, particularly rect and complicated interconnections, should be explored
in e-commerce, crime, healthcare, and the environment. to develop more viable solutions, policies, and strategies for
Between 2016 and 2018, security concepts, such as security mitigating surveillance vulnerabilities. Furthermore, addi-
systems, network security, and national security, gained pop- tional studies on the design and user experience of mobile
ularity, as did the employment of video surveillance, drones, applications may be beneficial to provide users with a greater

13
AI and Ethics (2023) 3:369–379 375

2012-2014: 2018-2020: After 2020:

Before 2012: 2014-2016: 2016-2018:
Development Data privacy, Public Health
Military and Societal security and
of machine regulations and Surveillance and
Border implications of video
learning and technology Transportation
Surveillance AI surveillance surveillance
algorithms acceptance Surveillance

Fig. 4  Historical evolution of concepts on AI surveillance

level of power and control over their privacy preferences. To the literature's flaws is the dearth of studies on citizens' per-
replace text-based, lengthy, and complex user agreements, spectives on the ethics of facial recognition and video sur-
interactive and simplified privacy agreements are required. veillance systems in transportation, law enforcement, and
Additionally, appropriate governance frameworks, ethi- national security. To address racial bias and inaccuracies in
cal norms, and rules for AI-based surveillance should be face recognition algorithms, more precise ethical regulations
devised to mitigate bias and socioeconomic inequity. Despite are expected, as these algorithms may result in prejudiced
the fact that the majority of governments adhere to their col- and discriminatory judgements, as well as severe effects on
laborations with private technology corporations, additional minorities and people of color. Additionally, there is the
strategies should be developed to foster civil society partici- matter of informed consent to consider. Several governments
pation, so they can better distinguish between state surveil- have deployed video surveillance cameras in public locations
lance of citizens and public health surveillance initiatives. and on streets. As these cameras currently capture individu-
A second significant worry voiced by social scientists als' facial data autonomously, additional means for obtaining
and humanities academics concerns the prominence of citizens' explicit and informed consent for the collection and
facial recognition and video surveillance technologies in the use of their biometric data must be devised.
transportation sector, as well as the blurred line between Military surveillance is the third topic that social scien-
this technology and national security. This issue is becom- tists and humanities scholars have addressed extensively.
ing increasingly significant as a result of the transportation The autonomy of military weapons, their accuracy and
industry's collaboration with police and national security precision in avoiding targeting non-combatants, and their
agencies to detect suspicious and terrorist activity. One of legal and political accountability are highly addressed by

13
376 AI and Ethics (2023) 3:369–379

Fig. 5  Varied perspectives regarding AI surveillance in healthcare, public transportation, military, pandemic management, urban planning, com-
munications, and big data

scholars. However, very few studies have examined interna- answer questions such as why certain features were selected
tional regulations and diplomatic tensions caused by military as features of a non-combatant civilian. In addition, more
drones. In addition, in contrast to existing studies that focus studies should be conducted regarding the partnership of
on the transparency of algorithms programming AI-based military agencies with technology companies that produce
machines, more research needs to be conducted on the "rea- advanced materials and fabrication technologies, as well
sonability" of algorithms to better understand the reasons as next-generation antennas, which increase the autonomy
behind how a drone decides who is a non-combatant civil- of drones, reduce the need for human operators, and allow
ian. The features and reasons for the engineering of military for collaborative autonomous information sharing among
weapons should be studied further. It should be clarified who drones. It is important to distinguish between the use of
decides and approves the algorithmic features and reasons drones for peaceful civilian purposes and the use of drones
that result in autonomous drone actions. Does an engineer for military and wartime purposes.
design features that are characteristic of a non-combatant One of the most prominent topics covered is the sur-
civilian and apply them to a drone? Do military stakehold- veillance of disease, the partnership of governments with
ers develop these features and deliver them to engineering technology companies, and the emergence of "surveillance
teams for use in developing algorithms and drones? In what capitalism," whereby companies monetize the data col-
manner do stakeholders determine the characteristics of a lected by tracking citizens' movements and behaviors. As a
civilian? In the event of an error, who is responsible? Was result of the development of mobile health apps embedded
it the result of human error? Did it occur due to a malfunc- with artificial intelligence and other digital technologies,
tioning machine? It would be beneficial to conduct further this topic expressed concerns about reducing the autonomy
research to understand the governance of "feature and reason and control of citizens over their movements and personal
engineering" of military drones, as well as the processes data. The studies focused primarily on the normative and
and procedures used during this process. While more AI societal ramifications of these applications on users' lives.
studies focus on the development of features, more stud- Research should be conducted to understand the strategies
ies should emphasize the development of reasons to better and alliances that technology companies utilize to convert

13
AI and Ethics (2023) 3:369–379 377

private user experiences into data-driven and predictive mar- 7 Conclusion

kets based on uncertain human futures. How should public
health startups, or companies such as Theranos and fake This study identified a series of the most often debated
patents be considered in this context? Following the govern- arguments around the surveillance effects of artificial intel-
ment's increase in funding for pandemic crises, a mushroom ligence. However, AI is not exclusively vulnerable to nega-
of startups and patents appeared, claiming their innovations tive monitoring mechanisms. Consequently, future research
and patents would provide efficient remedies. It is impera- might compare the detrimental and beneficial consequences
tive that more studies be performed on the functionalities of of AI surveillance. Moreover, in our study, we only con-
AI-based medical patents and innovations to determine the sidered the most frequently discussed controversial aspects
reliability and validity of innovations and distinguish them of AI surveillance. Future research may conduct system-
from fakes and fraudulent ones to reduce the monetary ben- atic literate reviews or text mining to identify alternative
efits of fraud and threats to human safety. scholarly discourses on AI surveillance. In this study, we
The fifth highly regarded topic is about smart cities concentrated on scholarly discussions of AI's disputed sur-
and the blurred boundaries between urban surveillance veillance implications. Future research can investigate the
and citizen ubiquitous surveillance in autonomous urban perspectives of other stakeholders in the AI ecosystem, such
environments. In urban contexts integrated with artificial as policymakers and citizens, to comprehend their opinions
intelligence, the Internet of Things, and ubiquitous com- on AI surveillance, for instance by conducting policy analy-
puting systems, there is a high level of contention related sis, Twitter analysis, or survey analysis.
to the autonomous collection of citizen data. In totalitarian
regimes as well as democratic regimes, the vertical trans-
mission of disparate and mass sources of urban data from Funding There is no funding for this research.
citizens into one entity raises concerns about panopticism
and persistent surveillance of citizens. To better understand
the ethical implications of embedded ubiquitous comput-
ing into an environment and context-aware services without References
citizens' awareness and control, more research is required
on informed consent, autonomy, and privacy. Most citizens 1. Almeida, D., Shmarko, K., Lomas, E.: The ethics of facial recog-
nition technologies, surveillance, and accountability in an age of
are unaware of what pieces of their personal information artificial intelligence: a comparative analysis of US, EU, and UK
are being captured, when they are captured, as well as with regulatory frameworks. AI Ethics 2021, 1–11 (2021). https://​doi.​
whom and for what purposes the captured information is org/​10.​1007/​S43681-​021-​00077-W
stored and shared. Moreover, human–computer interac- 2. Althar, R.R., Samanta, D.: The realist approach for evaluation of
computational intelligence in software engineering. Innov. Syst.
tion (HCI) scholars should conduct more design-centered Softw. Eng. (2021). https://​doi.​org/​10.​1007/​S11334-​020-​00383-2
research, so that the interface of context-aware services pro- 3. Ball, K., Di Domenico, M.L., Nunan, D.: Big Data Surveillance
vides citizens with a greater sense of autonomy and control and the Body-subject. Body Soc 22, 58–81 (2016). https://d​ oi.o​ rg/​
when interacting with autonomous environments. 10.​1177/​13570​34X15​624973
4. Bragazzi, N.L., Dai, H., Damiani, G., Behzadifar, M., Martini, M.,
The sixth highly debated topic pertains to computation Wu, J.: How Big Data and Artificial Intelligence Can Help Better
surveillance and soft computing as a subset of artificial Manage the COVID-19 Pandemic. Int. J. Environ. Res. Public
intelligence, fabrications, and fakeries that can be gener- Heal. 17, 3176 (2020). https://​doi.​org/​10.​3390/​IJERP​H1709​3176
ated through soft computing strategies. Computational 5. Brewster T, 2021. Government Secretly Orders Google To Iden-
tify Anyone Who Searched A Sexual Assault Victim’s Name,
intelligence has been extensively used to monitor civility in Address Or Telephone Number [WWW Document]. Forbes. URL
online communications. However, it has also been utilized https://w
​ ww.f​ orbes.c​ om/s​ ites/t​ homas​ brews​ ter/2​ 021/1​ 0/0​ 4/g​ oogle-​
to promote fakeness, such as deepfakes. Recent surpluses keywor​ d-w​ arran​ ts-g​ ive-u​ s-g​ overn​ ment-d​ ata-o​ n-s​ earch-u​ sers/?s​ h=​
of deepfakes necessitate more studies regarding the ethical 4d627​3c07c​97 (accessed 3.27.22).
6. Byrne, E.F.: Making drones to kill civilians: is it ethical? J. Bus.
challenges of deepfakes and the need for regulatory frame- Ethics. (2015). https://​doi.​org/​10.​1007/​S10551-​015-​2950-4
works that limit algorithm-based fabrications and manipu- 7. Caparini, M., Gogolewska, A.: Governance challenges of trans-
lations. Data surveillance and intrusive surveillance are the formative technologies. Connections 20, 91–100 (2021). https://​
last highly discussed topics. This topic addresses the issue doi.​org/​10.​11610/​CONNE​CTIONS.​20.1.​06
8. Chamikara, M.A.P., Bertok, P., Khalil, I., Liu, D., Camtepe, S.:
of invasion of privacy and security in the age of big data. In Privacy preserving face recognition utilizing differential privacy.
spite of this, very few studies have examined the topic of sur- Comput. Secur. 97, 101951 (2020). https://​doi.​org/​10.​1016/J.​
veillance and invasion in the age of big data. Furthermore, COSE.​2020.​101951
it is important to conduct more research to analyze aspects 9. Chang, V.: An ethical framework for big data and smart cities.
Technol. Forecast. Soc. Change 165, 120559 (2021). https://​doi.​
of intrusive and covert surveillance of citizens empowered org/​10.​1016/J.​TECHF​ORE.​2020.​120559
by data sources surrounding the everyday lives of citizens.

13
378 AI and Ethics (2023) 3:369–379

10. Cooke, P.: Image and reality: ‘digital twins’ in smart factory (2018). https://​doi.​org/​10.​1126/​SCIEN​CE.​AAO29​98/​SUPPL_​
automotive process innovation—critical issues. Reg. Stud. FILE/​AAO29​98_​LAZER_​SM.​PDF
(2021). https://​doi.​org/​10.​1080/​00343​404.​2021.​19595​44 28. Lin, P., Abney, K., Bekey, G.: Robot ethics: mapping the issues
11. Cosgrove, L., Karter, J.M., Morrill, Z., McGinley, M.: Psychol- for a mechanized world. Artif. Intell. 175, 942–949 (2011).
ogy and Surveillance Capitalism: The Risk of Pushing Mental https://​doi.​org/​10.​1016/J.​ARTINT.​2010.​11.​026
Health Apps During the COVID-19 Pandemic. J Human Psy- 29. Lou, W., Qiu, J.: Semantic information retrieval research based
chol 60, 611–625 (2020). https://​doi.​org/​10.​1177/​00221​67820​ on co-occurrence analysis. Online Inf. Rev. 38, 4–23 (2014).
937498 https://​doi.​org/​10.​1108/​OIR-​11-​2012-​0203/​FULL/​XML
12. Deshpande, A., Razmjooy, N., Estrela, V.V.: Introduction to 30. Mandal, V., Mussah, A.R., Jin, P., Adu-Gyamfi, Y.: Artificial
computational intelligence and super-resolution intell. Methods intelligence-enabled traffic monitoring system. Sustain 12, 9177
super-resolution image process. Appl Comput (2021). https://​ (2020). https://​doi.​org/​10.​3390/​SU122​19177
doi.​org/​10.​1007/​978-3-​030-​67921-7_1 31. Martinez-Balleste, A., Perez-Martinez, P., Solanas, A.: The pur-
13. Dewandre, N.: Big data: from modern fears to enlightened and suit of citizens’ privacy: a privacy-aware smart city is possible.
vigilant embrace of new beginnings. Big Data Soc (2020). IEEE Commun. Mag. 51, 136–141 (2013). https://​doi.​org/​10.​
https://​doi.​org/​10.​1177/​20539​51720​936708 1109/​MCOM.​2013.​65256​06
14. Dworzecki, J., Nowicka, I.: Artificial intelligence (AI) and 32. Maus, G.: Decoding, hacking, and optimizing societies: explor-
ICT-enhanced solutions in the activities of police formations ing potential applications of human data analytics in sociologi-
in Poland. Adv. Sci. Technol. Secur. Appl. (2021). https://​doi.​ cal engineering, both internally and as offensive weapons. Proc.
org/​10.​1007/​978-3-​030-​88972-2_​11 2015 Sci Inf. Conf. SAI 2015, 538–547 (2015). https://​doi.​org/​
15. Lage Dyndal, G., Arne Berntsen, T., Redse-Johansen, S.: Auton- 10.​1109/​SAI.​2015.​72371​95
omous military drones—no longer science fiction. Romanian 33. Miller, T.: Explanation in artificial intelligence: insights from
Military Thinking 2 (2017) the social sciences. Artif. Intell. 267, 1–38 (2019). https://​doi.​
16. Eck, K., Hatz, S.: State surveillance and the COVID-19 crisis. J org/​10.​1016/J.​ARTINT.​2018.​07.​007
Human Rights (2020). https://​doi.​org/​10.​1080/​14754​835.​2020.​ 34. Naudé, W.: Artificial intelligence vs COVID-19: limitations,
18161​63 constraints and pitfalls. AI Soc (2020). https://​doi.​org/​10.​1007/​
17. Feldstein, S., 2022. The Global Expansion of AI Surveillance S00146-​020-​00978-0
[WWW Document]. Carnegie Endow. Int. Peace. URL https://​ 35. Nguyen, M.T., Truong, L.H., Tran, T.T., Chien, C.F.: Artificial
carne​g ieen​d owme​n t.​o rg/​2 019/​0 9/​1 7/​global-​expan​s ion-​o f-​a i-​ intelligence based data processing algorithm for video sur-
surve​illan​ce-​pub-​79847 (accessed 3.10.22). veillance to empower industry 3.5. Comput. Ind. Eng. (2020).
18. Finn, R.L., Wright, D.: Unmanned aircraft systems: surveil- https://​doi.​org/​10.​1016/J.​CIE.​2020.​106671
lance, ethics and privacy in civil applications. Comput. Law 36. Oshikawa, R., Qian, J., Wang, W.Y.: A survey on natural lan-
Secur. Rev. 28, 184–194 (2012). https://​d oi.​o rg/​1 0.​1 016/J.​ guage processing for fake news detection. Lr. 2020—12th Int.
CLSR.​2012.​01.​005 Conf. Lang. Resour. Eval. Conf. Proc. (2018). https://​doi.​org/​
19. Gnatik, E.N., Hикoлaeвнa, ГE.: ‘New normality’ of the covid-19 10.​48550/​arxiv.​1811.​00770
era: Opportunities, limitations, risks. Rudn J. Sociol. 21, 769–782 37. Park, Y.J., Jones-Jang, S.M.: Surveillance, security, and AI as
(2021). https://​doi.​org/​10.​22363/​2313-​2272-​2021-​21-4-​769-​782 technological acceptance. AI Soc. 2021, 1–12 (2022). https://​
20. Guera, D., Delp, E.J., 2019. Deepfake Video Detection Using doi.​org/​10.​1007/​S00146-​021-​01331-9
Recurrent Neural Networks. Proc. AVSS 2018 - 2018 15th IEEE 38. Rajawat, A.S., Rawat, R., Barhanpurkar, K., Shaw, R.N., Ghosh,
Int. Conf. Adv. Video Signal-Based Surveill. https://​doi.​org/​10.​ A.: Vulnerability analysis at industrial internet of things plat-
1109/​AVSS.​2018.​86391​63 form on dark web network using computational intelligence.
21. Idrees, S.M., Nowostawski, M., Jameel, R.: Blockchain-based Stud. Comput. Intell. 950, 39–51 (2021). https://​d oi.​o rg/​1 0.​
digital contact tracing apps for COVID-19 pandemic manage- 1007/​978-​981-​16-​0407-2_4
ment: issues, challenges, solutions, and future directions. JMIR 39. Raza, K., Maryam, Q., S.,: An introduction to computational
Med Inf (2021). https://​doi.​org/​10.​2196/​25245 intelligence in COVID-19: surveillance, prevention, prediction,
22. Johnson, K.N., Reyes, C.L.: Exploring the implications of artifi- and diagnosis. Stud. Comput. Intell. 923, 3–18 (2021). https://​
cial intelligence. J. Int. Comp. L. 8, 315 (2021) doi.​org/​10.​1007/​978-​981-​15-​8534-0_1
23. Kalinin M.O., Krundyshev V.M.: Computational intelligence tech- 40. Ringrose, K., 2019. Law Enforcement’s Pairing of Facial Recog-
nologies stack for protecting the critical digital infrastructures nition Technology with Body-Worn Cameras Escalates Privacy
against security intrusions Proc 2021 5th World Conf. Secur. Sus- Concerns. Virginia Law Rev. Online 105.
tain Smart Trends Syst (2021). https://​doi.​org/​10.​1109/​WORLD​ 41. Rinik, C.: Data trusts: more data than trust? The perspective of
S4519​98.​2021.​95140​04 the data subject in the face of a growing problem. Int. Rev. Law
24. Kazansky, B.: ‘It depends on your threat model’: the anticipatory Comput. Technol. 34(3), 342–363 (2020). https://​doi.​org/​10.​
dimensions of resistance to data-driven surveillance. Big Data Soc 1080/​13600​869.​2019.​15946​21
(2021). https://​doi.​org/​10.​1177/​20539​51720​985557 42. Roussi, A.: Resisting the rise of facial recognition. Nature 587,
25. Kiliç, M., 2021. Ethico-Juridical Dimension of Artificial Intel- 350–353 (2020). https://​doi.​org/​10.​1038/​D41586-​020-​03188-2
ligence Application in the Combat to Covid-19 Pandemics 299– 43. Rubeis, G.: The disruptive power of artificial intelligence ethi-
317. https://​doi.​org/​10.​1007/​978-​981-​33-​6811-8_​16 cal aspects of gerontechnology in elderly care. Arch. Gerontol.
26. Lawlor, B.: An overview of the 2021 NISO plus conference: Geriatr. (2020). https://​doi.​org/​10.​1016/j.​archg​er.​2020.​104186
global connections and global conversations. Inf. Serv. Use 41, 44. Saheb, T., Amini, B., Kiaei Alamdari, F.: Quantitative analy-
1–37 (2021). https://​doi.​org/​10.​3233/​ISU-​210120 sis of the development of digital marketing field: Bibliometric
27. Lazer, D.M.J., Baum, M.A., Benkler, Y., Berinsky, A.J., Green- analysis and network mapping. Int. J. Inf. Manag. Data Insights
hill, K.M., Menczer, F., Metzger, M.J., Nyhan, B., Pennycook, G., (2021). https://​doi.​org/​10.​1016/j.​jjimei.​2021.​100018
Rothschild, D., Schudson, M., Sloman, S.A., Sunstein, C.R., Thor- 45. Saheb, T., Cabanillas, F.J.L., Higueras, E.: The risks and ben-
son, E.A., Watts, D.J., Zittrain, J.L.: The science of fake news: efits of Internet of Things (IoT) and their influence on smart-
addressing fake news requires a multidisciplinary effort. Science watch use. Spanish J. Mark. ESIC (2022). https://​doi.​org/​10.​
1108/​SJME-​07-​2021-​0129/​FULL/​PDF

13
AI and Ethics (2023) 3:369–379 379

46. Saheb, T., Saheb, M.: Analyzing and visualizing knowledge 60. Sylvia, J.J.: The biopolitics of social distancing. Media Soc Soc
structures of health informatics from 1974 to 2018: A bib- (2020). https://​doi.​org/​10.​1177/​20563​05120​947661
liometric and social network analysis. Healthc. Inform. Res. 61. Thilakarathne, N.N., Madhuka Priyashan, W.D.: An over-
(2019). https://​doi.​org/​10.​4258/​hir.​2019.​25.2.​61 view of security and privacy in smart cities. EAI/Springer
47. Saheb, T., Sabour, E., Qanbary, F., Saheb, T.: Delineating privacy Innov. Commun. Comput. (2022). https:// ​ d oi. ​ o rg/ ​ 1 0. ​ 1 007/​
aspects of COVID tracing applications embedded with proximity 978-3-​030-​82715-1_2
measurement technologies & digital technologies. Technol. Soc. 62. van Eck, N.J., Waltman, L.: Software survey: VOSviewer, a
69, 101968 (2022). https://​doi.​org/​10.​1016/J.​TECHS​OC.​2022.​ computer program for bibliometric mapping. Scientometrics 84,
101968 523–538 (2010). https://​doi.​org/​10.​1007/​s11192-​009-​0146-3
48. Saheb, T., Saheb, T., Carpenter, D.O.: Mapping research strands 63. van Zoonen, L.: Privacy concerns in smart cities. Gov. Inf. Q. 33,
of ethics of artificial intelligence in healthcare: a bibliometric and 472–480 (2016). https://​doi.​org/​10.​1016/J.​GIQ.​2016.​06.​004
content analysis. Comput. Biol. Med. 135, 104660 (2021). https://​ 64. Veloutsou, C., Mafe, C.R.: Brands as relationship builders in the
doi.​org/​10.​1016/J.​COMPB​IOMED.​2021.​104660 virtual world: a bibliometric analysis. Electron. Commer. Res.
49. Schmidt, A., Wiegand, M., 2017. A Survey on Hate Speech Appl. (2019). https://​doi.​org/​10.​1016/j.​elerap.​2019.​100901
Detection using Natural Language Processing. Soc. 2017 - 5th 65. Wai-Loon Ho, C., Caals, K., Zhang, H., Ho, L., Caals, K., Zhang,
Int. Work. Nat. Lang. Process. Soc. Media, Proc. Work. AFNLP H.: Heralding the digitalization of life in post-pandemic East
SIG Soc. 1–10. https://​doi.​org/​10.​18653/​V1/​W17-​1101 Asian Societies. J. Bioethical Inq. 174(17), 657–661 (2020).
50. Schoenherr, J.R., 2020. Understanding Surveillance Societies: https://​doi.​org/​10.​1007/​S11673-​020-​10050-7
Social Cognition and the Adoption of Surveillance Technologies. 66. White, C.L., Boatwright, B.: Social media ethics in the data econ-
Int. Symp. Technol. Soc. Proc. 2020-November, 346–357. https://​ omy: issues of social responsibility for using Facebook for public
doi.​org/​10.​1109/​ISTAS​50296.​2020.​94622​05 relations. Public Relat. Rev. (2020). https://​doi.​org/​10.​1016/J.​
51. Segun, S.T.: From machine ethics to computational ethics. AI Soc. PUBREV.​2020.​101980
36, 263–276 (2021). https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 00146-0​ 20-0​ 1010-1 67. Wright, N., 2018. How Artificial Intelligence Will Reshape the
52. Sekalala, S., Dagron, S., Forman, L., Mason Meier, B.: Analyzing Global Order. Foreign Aff. 10.
the human rights impact of increased digital public health surveil- 68. Yuan, M., Nikouei, S.Y., Fitwi, A., Chen, Y., Dong, Y., 2020.
lance during the COVID-19 Crisis. Heal. Hum. Rights J. 22, 7–20 Minor Privacy Protection through Real-time Video Processing at
(2020) the Edge. Proc. - Int. Conf. Comput. Commun. Networks, ICCCN
53. Shachar, C., Gerke, S., Adashi, E.Y.: AI surveillance during pan- 2020-August. https://​doi.​org/​10.​1109/​ICCCN​49398.​2020.​92096​
demics: ethical implementation imperatives. Hastings Cent. Rep. 32
50, 18–21 (2020). https://​doi.​org/​10.​1002/​hast.​1125 69. Završnik, A.: Introduction: Situating Drones in Surveillance
54. Shimizu, Y., Osaki, S., Hashimoto, T., Karasawa, K.: How do Societies. Drones Unmanned Aer. Syst. Leg. Soc. Implic. Secur.
people view various kinds of smart city services? Focus on the Surveill (2016). https://​doi.​org/​10.​1007/​978-3-​319-​23760-2_1
acquisition of personal information. Sustain. 13, 11062 (2021). 70. Zhang, K., Ni, J., Yang, K., Liang, X., Ren, J., Shen, X.S.: Secu-
https://​doi.​org/​10.​3390/​SU131​911062 rity and privacy in smart city applications: challenges and solu-
55. Shook, J.R., Solymosi, T., Giordano, J.: Ethical constraints and tions. IEEE Commun. Mag. 55, 122–129 (2017). https://​doi.​org/​
contexts of artificial intelligent systems in national security, intel- 10.​1109/​MCOM.​2017.​16002​67CM
ligence, and defense/military operations. Artif. Intell. Glob. Secur. 71. Zhang, Z., Zaman, A., Xu, J., Liu, X.: Artificial intelligence-aided
(2020). https://​doi.​org/​10.​1108/​978-1-​78973-​811-​72020​1008 railroad trespassing detection and data analytics: methodology and
56. Smith, B.W., 2020. Ethics of Artificial Intelligence in Transport, a case study. Accid. Anal. Prev. (2022). https://​doi.​org/​10.​1016/J.​
in: Dubber, M., Pasquale, F.,, Das, S. (Eds.), The Oxford Hand- AAP.​2022.​106594
book of Ethics of Artificial Intelligence. 72. Ziosi, M., Hewitt, B., Juneja, P., Taddeo, M., Floridi, L.: Smart cit-
57. Smith, M., Miller, S.: The ethical application of biometric facial ies: mapping their ethical implications. SSRN Electron J (2022).
recognition technology. AI Soc (2021). https://​doi.​org/​10.​1007/​ https://​doi.​org/​10.​2139/​SSRN.​40017​61
S00146-​021-​01199-9
58. Suchman, L.: Algorithmic warfare and the reinvention of accu- Publisher's Note Springer Nature remains neutral with regard to
racy. Crit. Stud. Secur. 8(2), 175–187 (2020). https://​doi.​org/​10.​ jurisdictional claims in published maps and institutional affiliations.
1080/​21624​887.​2020.​17605​87
59. Suratkar, S., Bhiungade, S., Pitale, J., Soni, K., Badgujar, T., Kazi,
F.: Deep-fake video detection approaches using convolutional –
recurrent neural networks. J. Control Decis. (2022). https://​doi.​
org/​10.​1080/​23307​706.​2022.​20336​44

13