Children’s Behavior on the Internet: Conceptualizing the

Synergy of Privacy and Information Disclosure

Aseel Alghafis Ali Alkhalifah
Department of Information TECHNOLOGY, College of Department of Information TECHNOLOGY, College of
Computer, Qassim University Computer, Qassim University

ABSTRACT social networking sites [3]. However, despite the interest in chil-
Children’s behavior on the Internet and their information disclo- dren’s behavior on the Internet, little research has surprisingly been
sure and privacy are essential topics in academic research and conducted on these topics compared to other issues, even though
practice. The main focus of this study assesses current research children’s online behavior and information privacy are today con-
concerning children’s online behavior related to privacy and in- sidered extremely important.
formation disclosure and analyzes the extant studies. Furthermore, In this study, the three layers of the technical–formal–informal
the study sheds light on the emergence and present state of these (TFI) model are used to analyze prior studies [4],[5]. The current
concepts through a thorough analysis and classification of papers in study explains what is meant by privacy, information disclosure,
this domain from reputable researches obtained from the literature. children’s online behavior, and privacy with machine learning and
Thus, this study establishes a framework for categorizing research, reviews studies on these topics. Moreover, statistical techniques
summarizing significant findings, and highlights research gaps and and procedures are used to investigate and determine the extent
future opportunities concerning children’s online behaviors and of the Internet’s effects on children. Hence, this study focuses on
activities. reviewing the potential impact of children’s online behavior, which
has become one of the most critical issues in today’s literature.
CCS CONCEPTS
2 LITERATURE REVIEW
• Information systems; • World Wide Web; • Web applica-
tions; 2.1 Research Method
The synthesis of significant themes and issues concerning this study
KEYWORDS and the organization and classification of existing studies have
Privacy, Children, Behavior, Information Disclosure given us an overview of the considerable research areas regarding
the predefined category. This has been possible through the use of a
ACM Reference Format: structured literature review (SLR) [4]. The SLR is mainly viewed as
Aseel Alghafis and Ali Alkhalifah. 2021. Children’s Behavior on the Internet:
an ideal method that will allow the reviewer to cover broader scope
Conceptualizing the Synergy of Privacy and Information Disclosure. In
and range of literature on children’s online behavior. Thus, this
The 2021 9th International Conference on Computer and Communications
Management (ICCCM ’21), July 16–18, 2021, Singapore, Singapore. ACM, New study adopted the review procedure recommendations suggested
York, NY, USA, 7 pages. https://doi.org/10.1145/3479162.3479180 by Alkhalifah and D’Ambra [4].
Relevant papers on children’s behavior on the Internet are down-
loaded from two databases: Google Schooler search IEEEexplore
1 INTRODUCTION digital database. Moreover, several keywords were identified and
The Internet’s primary purpose is evolving toward sociability in- applied during the search process, which comprises "children’s be-
stead of information retrieval, while the Internet age has seen the havior on Internet," "privacy," "information disclosure," and "risk."
rise of websites and games in addition to social networking [1]). A Finally, the literature evaluation was restricted to the period from
new risk is thus arising for adolescents due to the emergence of January 2007 to December 2018. The latter date indicated the con-
online activities [2]. Recently, the population of children engaging clusion of this research project, and the former date marking the
in online activities is steadily increasing, surpassing the adults’ beginning of the study.
usage rates [3]. For instance, on an average day, seven out of 10 The study undertook several steps. Papers were firstly collected
children go online [3]. Furthermore, due to significant online access by reading their titles and abstracts. Articles not relevant to chil-
worldwide for online games, most children are now accessing the dren’s behavior on the Internet concerning privacy, information
Internet regularly and engaged in online chat and visiting other disclosure, or risks were excluded. Then, each paper was classified
according to whether it was a book/book chapter, journal arti-
Permission to make digital or hard copies of all or part of this work for personal or cle, technical report/white paper, or conference proceedings, and
classroom use is granted without fee provided that copies are not made or distributed
for profit or commercial advantage and that copies bear this notice and the full citation whether it was based on the stated exclusion criteria. A paper was
on the first page. Copyrights for components of this work owned by others than ACM selected if it met the inclusion criteria or was eliminated if it met
must be honored. Abstracting with credit is permitted. To copy otherwise, or republish,
to post on servers or to redistribute to lists, requires prior specific permission and/or a
any exclusion criteria (see Table 1).
fee. Request permissions from [email protected].
ICCCM ’21, July 16–18, 2021, Singapore, Singapore 2.2 Technical–Formal–Informal (TFI) Model
© 2021 Association for Computing Machinery.
ACM ISBN 978-1-4503-9007-1/21/07. . . $15.00 An established tradition has developed in the computer science
https://doi.org/10.1145/3479162.3479180 (CS) and information systems (IS) domain in which the research

117
ICCCM ’21, July 16–18, 2021, Singapore, Singapore Aseel Alghafis and Ali Alkhalifah

Table 1: Research Classification

Research Sources Number %

Journal Articles 23 %92
Conference Papers or Proceedings 2 %8
Reports (Technical or White Papers) 0 0
Book Chapters 0 0
Total 25 %100

literature itself is evaluated to understand better the current posi-

tion about the field of research interest and to distinguish current
patterns, focus, and advancement of the discipline [4]. In this re-
gard, the primary goal of this study was to conduct a review of the
literature to gain a better understanding of children’s behaviors
and activities on the Internet.
It is beneficial to use a conceptual framework to categorize ear-
lier research papers on a topic when analyzing prior research [4].
There a study [4] that made use of a technical–formal–informal
(TFI) framework as shown in 1, which conceptualized and divided
IS research into three distinct but interconnected layers (technical Figure 1: Technical–formal–informal (TFI) model of infor-
(T), formal (F), and informal (I)). The TFI framework was used in mation systems [5]
this study. The TFI layers provide a more holistic approach to ana-
lyzing the related themes. Hence, the layer-specific to the research
themes that are related are easily understood and categorized. This 2.3 Results
helps in placing the research according to the context that could When framing and categorizing the literature in this study, the TFI
belong [5]. Furthermore, the TFI model can identify research gaps model was used, previously introduced after identifying the various
and conclude the consequences of future research and practices perspectives of research concerning children’s behavior on the
[4].Nonetheless, the literature has provided us with knowledge Internet used in the previous study [4]. The analysis focused on the
about the basic definitions used to explain the three layers of the relationship between each of the selected papers and the technical,
TFI model. Firstly, the technical layer is the parts of formal systems formal, and informal layers. After completing the classification
that focus on automation and standardization (computers providing process, we discovered that 3 studies (12%) were technical, 11
help to operational tasks). The formal covers governance, policies, studies (44%) investigations were formal, and 11 studies (44%)
or processes, including organizational-level security mechanisms studies were informal [4] (see Fig. 2). The findings demonstrated
(developing controls used in the organizational structure). Finally, that a reasonable amount of study had been conducted on formal
the informal covers shaping employees’ beliefs, norms, attitudes, and informal levels, with the formal and informal layers taking
and values, including individual-level security mechanisms, like the lead in terms of importance. The technological research
establishing normative controls [5]. Moreover, the layers of the TFI layer, on the other hand, has received little attention. This
model are commonly considered interdependent [4]. link (https://docs.google.com/document/d/e/2PACX-1vR6vq80-J-
The interrelationships between these abstracted layers are de- skz6KmJCTnLGn5PRtCqZHQpqlq0B1Eh2DwrZxvFTKDF3E11NPVj
scribed as follows [5]: "because formal norms can only operate by Fq4Dq7uork7COY613U/pub) provides a description of the pub-
the informal norms needed to interpret them, informal norms are lications identified in the selection phase of the literature
fundamental, while technical norms can play no role ... unless em- review.
bedded within a system of formal norms" [5]. In other words, these
interrelationships cannot permanently be embedded in technical 3 PRIVACY
systems or replaced by rules, especially considering the informal
means by which organizations managed information as a critical or- There are various sub-definitions for the term "privacy." Individual
ganizational resource [5]. However, the relationships between these privacy can range from "personal privacy" to "information privacy,"
three levels become even more complex when information systems with the latter being a phrase for privacy that refers to all of the
(IS) move across single organization boundaries (i.e., interoperable different ways in which it can be defined concerning the Internet
systems and virtual organizations): when this occurs, additional [6]. Although multiple authors have presented distinct definitions
issues, comprising privacy and trust based on new informal and for privacy and information privacy [6], there is no consensus on
formal mechanisms, need to be addressed [5]. what constitutes privacy and information privacy. Moreover, no
specific description of privacy covers all aspects of the term, with
the concept of privacy understood in varied ways [6]. One of the
most significant "ethical challenges of the digital era" is protecting
personal information [7]. Multiple disciplines in the social sciences,

118
Children’s Behavior on the Internet: Conceptualizing the Synergy of Privacy and Information Disclosure ICCCM ’21, July 16–18, 2021, Singapore, Singapore

Figure 2: Classification of research papers based on TFI layers

ranging across psychological, philosophical, sociological, and legal profiles that included an image, and 50.8% had put their present
approaches, have extensively studied the concept of privacy [7]. home as their current location. Aside from that, 39.9% of users had
Many concerns have been raised about the spread of children’s provided a phone number, and most users had provided their full
Internet use and the disclosure of their privacy [8]. names [10]. Online privacy disclosure was also observed in other
Generally, the term "information privacy" refers to the unau- socially based Internet domains, such as forums and blogs. One
thorized collection of information, its illegal use, inaccuracies and such medium is an online blog, which is an online journal [10]. A
errors, and improper access to individuals’ control over when their little more than half (54%) of the users in the survey had explicitly
personal information is released [9]. The right to be left alone has provided demographic information, such as their occupation, age,
been mainly considered synonymous with privacy [7]. However, and geographic area, whereas 92%t of blog authors had explicitly
many have claimed that privacy refers simply to the right to pro- provided their correct name [10]. This is important because kids
hibit disclosing personal information to third parties [9] and that no dominate such blogs, with 52% of all blogs written and maintained
further rights exist. Reference [8] defines privacy as a right that al- by teens between 13 and 19 years [10]. When a person P’s private
lows users to maintain control over their disclosures, retain control information is disclosed by their friend FP to another person FFP
over their data, and be left alone with their thoughts and feelings. (who is not P’s friend), this is referred to as "privacy leakage" in on-
The term "privacy" refers to the "interest individuals have in main- line social networks [11]. Reference [14] stated that privacy leakage
taining personal space free from interference by other persons and might continue to occur throughout the entire time that personal
organizations." information is shared through engagement with friends. Some in-
Further, reference [11] explains Information privacy as an in- dividuals interested in obtaining a user’s information may attempt
dividual’s desire to have control or influence over their personal to gain access to that information in any way possible (legally or
information. As a result of sharing content (such as images, videos, illegally). Refusing to give up may lead to severe concerns [14].
and locations) with others, most people’s personal information may Despite the high level of concern among users, according to
be exposed to unauthorized access or those with harmful intent if a study conducted on five Russian social network service (SNS)
the information is not properly controlled. As a result, the problem providers [15], there is a lack of understanding of privacy-related
of privacy is fundamental [12]. Teenagers and children are classified ideas and the necessity of preparing appropriate defensive mecha-
under the demographic that is particularly prone to privacy viola- nisms for users. It is generally considered to be devastating when
tions. This is because they typically reveal personal information a vast amount of users’ information is exposed. A significant gap
in public and are more susceptible to being fooled by fraudulent has been discovered between these providers and their Western
users. Therefore, they may be more easily influenced to share their counterparts regarding privacy protections [15]. These automated
information [12]. There is also a privacy issue concerning the self- systems can perform activities such as creating fictitious profiles,
disclosure of personal information to consider [13]. Self-disclosure crawling the Internet, and sending falsified friendship requests [15].
is the process by which an individual reveals information about The concept of privacy risk has been described by several authors
themselves to other individuals [13]. [3, 16]. This refers to the information disclosure of personal data
Two characteristics of information appear to be crucial to in- due to misuse of personal information and opportunistic conduct.
formation disclosure: intimacy and sensitivity [13] and privacy Hence this behavior is associated with internet users’ anticipation
disclosure of the amount and type of information that individuals of financial losses. Furthermore, when the losses produced by the
are willing to disclose to another [14]. According to one survey personal information disclosure are more than the risks, users will
[10], 87.8% of users had provided their birth date, around 90.8% had perceive the risk as being greater [16].

119
ICCCM ’21, July 16–18, 2021, Singapore, Singapore Aseel Alghafis and Ali Alkhalifah

Many studies have proven that there is a relationship between other aspects. Traditional definitions of self-disclosure [20] include
privacy concerns and privacy risks to protect personal informa- situations in which a person reveals their own identity to another.
tion. The authors of [16] showed evidence of a positive relationship Aside from giving personally identifiable information, users of on-
between privacy risk and personal information misuse without sug- line social networks (OSNs) can expose various private information
gesting a causation relationship between the two. In a process de- on their accounts, such as their musical preferences, hobbies, or
signed to assist individuals in protecting their personal information favorite movies [20]. In addition, users frequently submit their im-
from unauthorized access and violators, it has been demonstrated ages and disseminate the news through comments or postings on
that Internet users’ privacy concerns positively affect privacy risk the "Wall" [20], which is becoming increasingly popular.
[16], with evidence of this relationship also provided in [17]. Re- In the relationship between disclosure practices and privacy atti-
duced perceived risks were shown to enhance the likelihood of tudes, many variables are considered [21]. One of these variables is
dangerous activities, this evidence according to the research on the composition of one’s sharing network [21], with this particularly
the privacy of children’s information [3]. In this case, disclosing underlining the importance of self-similarity in friendship interac-
information in a way compatible with the amount of children’s tions on social networking sites. Two elements, namely, privacy
privacy concerns was likely because privacy concerns encouraged policy and privacy settings, can be directly influenced by a social
risk-reducing activities. Reference [3] reported that as children’s networking site [21]. Among the elements influencing interaction
privacy worries grew, their willingness to provide personal infor- with "strangers" in an online environment and personal information
mation reduced. disclosure [22] are the characteristics of mediated communication.
Two commonly used factors related to information can be used The need to maintain one’s social standing within peer networks
as inputs to measure users’ privacy disclosure scores and users’ and the public character of contacts and the desire to gain social
privacy on online social networks [6]—These factors are comprised approval may impact risk-taking behavior, and the information
of visibility and sensitivity. The sensitivity of a user’s information exposed [22], among other things. Young people may reveal sensi-
refers to the process by which the risk posed by the disclosure of tive data about themselves more quickly due to their heightened
the user’s information grows and the increased risk associated with vulnerability to manipulation by persons they encounter online
the user’s attributes when that attribute is sensitive. For example, inside their current peer networks [22]. To get the best possible
a study by [6] computed sensitivity scores for 11 privacy-related level of privacy, we must detect or be aware of the scope of our
variables based on the data collected from online users. The find- disclosures [22].
ings revealed that the most sensitive qualities were associated with From a practical standpoint, this entails knowing the boundary
religious beliefs, political beliefs, relationship status, and phone of a room by assuming that others outside cannot hear and the
numbers. However, users’ did not regard two attributes to be dis- trustworthiness of a communication partner to know who will gos-
creet. This includes the date of birth and the location of their present sip and who will not. This is in addition to understanding the data
residence. Further, the visibility factor is used to examine when a practices of the website [21]. The relationship between information
user’s traits on an online social network are evaluated to determine exposure and privacy attitudes has been extensively researched [21].
their level of accessibility to other users. Furthermore, the work Individuals would strive to balance the necessary disclosures of
by [6] took into account the ease with which a user’s information communication and the control mechanisms of individual privacy,
could be accessed, the difficulty with which data could be extracted, as this is a general optimization function. It has been discovered
and the frequency with which information was disclosed (data reli- that information disclosure is innately related to practical tools that
ability) to calculate the visibility of a user’s information in online change information disclosures following the individual’s goals,
social networks. knowledge, and privacy attitudes. Hence, the social individual must
know and choose when and where to disclose personal information
[21].
4 INFORMATION DISCLOSURE The work by reference [10] conducted a study in which they
In the context of individual users, information disclosure refers discovered the "information revelation" tendencies of over 4,000
to the extent and quantity of sensitive information that they have university students who used Facebook. They confirmed that users’
disclosed about themselves [18]. For greater specificity, information broad interests, articulated as category-based representations, are
disclosure is a composite variable that is determined by the sensitiv- common among most social networking sites [10]. Literary and
ity and quantity of information disclosed by the users. Examples of entertainment interests, as well as political orientation, are all in-
these instances include a prerequisite disclosure that is frequently cluded in this category. Furthermore, from discoveries about a user’s
required to make online purchases, personalize these services in inner life or intimate portrayals of their social activities, the user
the form of suggestions or "one-click" purchasing, or access online frequently provides identifiable data.
services via ubiquitous registration form [19]. Although, there are tools available on social networking sites
The increasingly social nature has put an associated privacy cost that enable individuals to establish their norms for disclosure [21].
on internet users. This is seen by both the disclosure of personal For instance, Facebook offers various privacy-related tools such as
information and the increased obligation to work correctly for the privacy management, friend groupings, item-level access control,
system [19]. Users must make a variety of privacy-related decisions and blocklists. Furthermore, there are also additional strategies for
concerning levels of public access, based on personal risk and secu- maintaining privacy. For example, users can enable the design of
rity factors and other considerations, such as compelling users to deterministic rules that define where content should be shared or re-
submit location-aware images taken with camera phones [19] and produced and which content can be used for various circumstances.

120
Children’s Behavior on the Internet: Conceptualizing the Synergy of Privacy and Information Disclosure ICCCM ’21, July 16–18, 2021, Singapore, Singapore

Interestingly, recent studies [21] have discovered an increased use the digital age. As a result, we are now witnessing the emergence
of privacy settings on computers and mobile devices. of a new childhood period in the digital world.
A growing amount of research has studied the consequences of As evidenced by recent literature [24], researchers have been
the users’ prevalence involvement in these kinds of behaviors [22], particularly interested in the consequences of Internet usage. Inves-
and the findings have been published in peer-reviewed journals. tigations examining the possible benefits from an emotional and
According to a survey conducted in the United Kingdom (UK), 29% psychosocial standpoint are being conducted. However, the detri-
of respondents, aged 11–16 years, have engaged with someone mental psychological impacts of the Internet on young people and
online with whom they had no prior contact. Furthermore, 36% their online danger continue to be a source of concern. According to
had accepted friend requests from people they had never met in the findings of [24], 93% of children aged 12–17 years in the United
person, and 14% had revealed personal information, such as ad- States (US) used the Internet in 2009, while 60% of a pan-European
dress or phone number, to people they had met online. Also, other sample of young people (aged 9–16) were demonstrated to be online
studies have indicated high prevalence rates of cyberbullying, with virtually every day in the same year. The rising online presence of
reference [22] reporting a rate of 28%. Many young people and youths and children has increased concerns regarding the suscep-
adults have utilized their personal information in their online ac- tibility of young customers, particularly from the perspectives of
counts as proof of inclusion to lessen their anonymity and establish privacy and security [3].
their identities [22], which has helped them become more visible. A study conducted by [23] shows that 75%% of respondents aged
Accordingly, in the UK, only 7% of the participants are between between 9–19 years had domestic Internet access, 37% of those
the ages of 9 and 16 years old had lied in their online profile. This aged 5–6 years, and 64% of those between the ages 7 8 years. 67%
shows that most young people use their online profiles to transmit of respondents aged between 8–10 years had never been alone to
personal information to establish their identity, rather than taking the grocery store or the local town in contracts. In total, 92% of
advantage of the anonymity afforded by the online environment to the respondent had access to the Internet through their schools,
participate in dishonest behavior [22]. with at least 84% reporting that they were regular Internet users.
Note that privacy in social networks is not solely determined by As a result, while society has made attempts to protect children
technical rules [21], which is a point to consider. In the study by from physical dangers externally, less attention has been given to
[21] entitled "contextual privacy on Facebook," it was discovered the children’s interactions with the technology [23]. Hence, the
that users exert control over their disclosures through the use of number of interactions has been increasing [3],[4],[22],[23]. A well-
both "mental" and "technological" measures to conceal their iden- known expert from the computer and artificial intelligence (AI) field
tities. The research demonstrated a complex connection between observed a passionate affair between youngsters and computers
elements connected to attitudes that influence privacy strategies throughout the world... they appear to understand deeply that
and the usage of privacy settings, which are the technical techniques computer technology belongs to them. They are well aware that
of disclosure control. The most challenging problem in modeling they belong to the "computer generation [25].
mental methods is identifying them. According to reference [21], Information collection approaches and tactics that make use of
even though these methods are common, it isn’t easy to reliably acquisition technologies and complex data mining have become
measure them. more user-friendly in recent years [3, 4]. According to research
by [3], data collecting occurs when children interact with a fic-
tional marketing character, offer personal information to social
5 CHILDREN’S BEHAVIOR ON THE networking sites, and register for websites or contests. According
INTERNET to reference [3], 25% of children’s websites gave memberships to
Firstly, the Internet is constantly being utilized by a technologically children under 12 years, while 73% of children’s websites utilized
sophisticated generation of adults and children to network and com- online games. Because the children are typically innocent, they
municate: it has become a feature of contemporary life and a vehicle have become easy prey for information gatherers [3, 4].
for the transmission of information [23]. As a result of the rapid The impact of the Internet on children is an essential issue to
growth of technology, significant social change can be observed be examined, and the following factors are only a few of the many
in virtually every aspect of contemporary society. However, since that affect children:
children and adults can connect through the same online facilities,
children are put at danger since they are exposed to inappropriate
content for someone of their age [23]. 5.1 Cognitive and academic impact.
In many countries, the rate of Internet usage among young con- The majority of parents believe that the Internet has a beneficial
sumers outpaces that of adults, resulting in a significant increase effect on their children’s academic progress, with 81% believing
in the number of youngsters worldwide who have access to the that the Internet is a good place for children to learn and 84%
Internet [3]. The children use the Internet to check their emails, believing that the Internet is helpful for home assignments [25]. In
chat online, do school assignments, use social networking sites, and fact, according to the parents, the most prevalent reason for their
play online games available in many languages [3] in places with children to use the Internet is to complete schoolwork.
online access. For example, in 1981, less than 300 computers could Social impact of the Internet: Parents should assess if the Internet
connect to the Internet in the UK. But that figure has now climbed is good or damaging to their children’s participation in their com-
to an uncountable level [23]. Moreover, in the past two decades, munities and interpersonal connections [25] before allowing their
there has been a total revolution in technology and a shift towards children to use it. Children’s rising isolation is attributed to Internet

121
ICCCM ’21, July 16–18, 2021, Singapore, Singapore Aseel Alghafis and Ali Alkhalifah

use. The study by the Stanford Institute for the Qualitative Study of protect privacy, the system that generates information about an
Society conducted with more than 4,000 parents in 1999 supported underlying dataset must be resistant to changes. Also, to better
this claim about rising isolation due to internet usage by children understand the dangers, attacks, and defenses faced by systems
[25]. A significant proportion of those who said they spent more built on machine learning, which provides automated methods for
time online also said they spent less time with people in the real processing massive amounts of data, numerous actions have been
world. Eight percent (8%) of the participants reported decreased carried out [28].
social activities, with around 25% reporting less time spent with The tasks that are solved using machine learning techniques are
their family or friends on the phone or in-person if they spent five typically divided into three types: the structure of the data analyzed
hours or more per week online. by the algorithm [28]. The first type is supervised learning, repre-
sented by the structure of the data analyzed by the algorithm. The
5.2 Physical impact of the Internet learning procedures are developed as a result of supervision. They
involve approaches based on training examples used to establish
With the Internet invading children’s lives daily, parents and others
a relationship using inputs labeled with equivalent outputs. If the
must notice the Internet’s harmful effects on children’s mental and
information comes from output categories, the task is referred to as
physical health [25]. There have been numerous concerns concern-
classification, and the domains derived from the real-valued output
ing the passive nature of computer use as a contributing factor
are used to determine the regression problems. Object recognition
to developing the childhood obesity issue in Western society. In
in machine translation, pictures, and spam filtering are examples of
a study conducted in the US, it was shown that more than 15% of
jobs classified as supervised learning [28]. In unsupervised learning,
children aged between 6–19 years were overweight in 1999–2000,
the method is given inputs that have not been labeled, and the job
which was more than triple the percentage seen in the 1960s. Inac-
is completed without using a label. The challenges considered in
tivity appears to be one of the most significant contributing factors
unsupervised learning involve dimensionality reduction to cluster-
to obesity in young people. A youngster spends an average of five
ing points. According to a similarity metric and projecting data in
hours per day on a computer, watching television, or doing home-
lower dimensional subspaces, this is done in the model pre-training
work, resulting in a relatively inactive child for 76 percent of the
[28]. For example, clustering can be used to find anomalies in data.
day [25, 26].
Among the approaches that fall under the third category of ma-
From the standpoint of children’s safety, it is essential not to
chine learning, methods that learn a policy for action over time
disregard the issue of direct contact with persons they encounter
through sequences of actions, observations, and rewards are in-
on social media platforms [26]. Youngsters must be made aware of
cluded [28]. As a discipline within machine learning, reinforcement
the possibility of manipulation and the risk that may arise when
learning can be considered the branch of that study that deals with
they come into contact with a manipulative individual. Moreover,
planning and control. Hence, the use of reinforcement learning
they must be made aware of the measures that must be taken in
in conjunction with unsupervised and supervised approaches was
this regard [26]. [26]
demonstrated to be effective in defeating humans [28].
Many researchers have been interested in children’s behav-
6 EMERGING TOPICS ior on the Internet [3],[23],[24],[25],[26]. Some have studied this
By using example data or prior experience, machine learning ex- phenomenon from the privacy perspective [6],[7],[8],[9],[10],[11],
plains the process of programming computers so that a performance [12],[13],[14],[15],[16], while others have researched the aspect
criterion is maximized using example data or previous experience of information disclosure [10],[18],[19],[20],[21], and others again
[27]. As a result of its tremendous ability to analyze vast volumes have explored privacy in relation to a specific topic, such as ma-
of data, machine learning has found widespread use in various chine learning [27],[28],[29]. The current study has summarized
industries in recent years. Computer-aided design (CAD) systems the following research gaps based on the literature review:
are utilized to address issues in a wide range of fields, including
• Few studies have focused on children’s behavior on the In-
classification, regression, and policy learning [28].
ternet.
Machine learning aims to get usable knowledge from large
• Few studies have investigated privacy concerns and their
amounts of data, such as classification algorithms, methods for
effect on information sharing.
forecasting quantities, and methods for discovering clusters of sim-
• Few studies on privacy have focused on a specific field, such
ilar samples [29]. Machine learning algorithms are used to select
as using machine learning techniques.
from a group of learning models depending on specific input data
and then provide the best output based on that selection. The out-
put model can be used to deal with future data or to analyze the 7 CONCLUSION
distribution of the data, depending on the situation. Therefore, it This study has provided a general overview of the current state of
is necessary for the output model to capture some information research on children’s behavior on the Internet from the categoriza-
about the dataset, even if the underlying dataset is often much tion and analysis conducted via the statistical method of previous
more compact than the output model. Within machine learning, investigations concerning information privacy and disclosure. The
there is a widespread demand for privacy and a growing interest study covers computer science, information systems/information
in the science of security concerns, which has not been ignored technology (IS/IT), and social sciences. Twenty-five (25) articles
[29]. Privacy refers to the act of preventing private data from be- have been classified using the technical–formal–informal (TFI)
ing exposed, mainly personal information about persons [29]. To model. A research framework has been developed based on all

122
Children’s Behavior on the Internet: Conceptualizing the Synergy of Privacy and Information Disclosure ICCCM ’21, July 16–18, 2021, Singapore, Singapore

potential relationships between these three separate layers: tech- [11] N. K. Lankton, D. H. McKnight, and J. F. Tripp, "Facebook privacy management
nical, formal, and informal. However, the number of studies was strategies: A cluster analysis of user privacy behaviors," Computers in Human
Behavior, vol. 76, pp. 149-163, 2017.
limited as papers were retrieved from only two databases. Future re- [12] C. S. Silva, G. A. Barbosa, I. S. Silva, T. S. Silva, F. Mourão, and F. Coutinho, "Privacy
search could be conducted with more papers from various databases for Children and Teenagers on Social Networks from a Usability Perspective:
A Case Study on Facebook," in Proceedings of the 2017 ACM on Web Science
in different fields. Conference, 2017, pp. 63-71: ACM.
Few studies have focused on design or technical problems and [13] K. Li, X. Wang, K. Li, and J. Che, "Information privacy disclosure on social
issues and challenges of children’s online behavior and their pri- network sites: An empirical investigation from social exchange perspective,"
Nankai Business Review International, vol. 7, no. 3, pp. 282-300, 2016.
vacy. Moreover, the perceptions of disclosing their identity from [14] X. Zhang, L. Zhang, and C. Gu, "Security Risk Estimation of Social Network
children’s viewpoints have not been carefully investigated in the ex- Privacy Issue," in Proceedings of the 2017 the 7th International Conference on
isting researches. Hence, this study found that suitable frameworks Communication and Network Security, 2017, pp. 81-85: ACM.
[15] M. Soryani and B. Minaei, "Social networks research aspects: A vast and fast
or models to understand better children’s online behavior are lim- survey focused on the issue of privacy in social network sites," arXiv preprint
ited. Therefore, more studies on the technical challenges and further arXiv:1201.3745, 2012.
[16] F. Xu, K. Michael, and X. Chen, "Factors affecting privacy disclosure on social
investigation of children’s perceptions and their requirements are network sites: an integrated model," Electronic Commerce Research, vol. 13, no.
essential to address the research gap between information theory 2, pp. 151-168, 2013.
and practices. [17] Y. Jordaan and G. Van Heerden, "Online privacy-related predictors of Facebook
usage intensity," Computers in Human Behavior, vol. 70, pp. 90-96, 2017.
[18] Z. De Souza and G. N. Dick, "Disclosure of information by children in social
REFERENCES networking—Not just a case of “you show me yours and I’ll show you mine”,"
International Journal of Information Management, vol. 29, no. 4, pp. 255-261,
[1] S. Livingstone and D. R. Brake, "On the rapid rise of social networking sites: New
2009.
findings and policy implications," Children & Society, vol. 24, no. 1, pp. 75-83,
[19] N. Joinson, U.-D. Reips, T. Buchanan, and C. B. P. Schofield, "Privacy, trust, and
2010.
self-disclosure online," Human–Computer Interaction, vol. 25, no. 1, pp. 1-24,
[2] J. Y. Kim, J. S. Lee, and S. Oh, "A path model of school violence perpetration:
2010.
introducing online game addiction as a new risk factor," Journal of Interpersonal
[20] H. Krasnova, S. Spiekermann, K. Koroleva, and T. Hildebrand, "Online social
Violence, vol. 32, no. 21, pp. 3205-3225, 2017.
networks: Why we disclose," Journal of Information Technology, vol. 25, no. 2,
[3] M. O. Lwin, A. D. Miyazaki, A. J. Stanaland, and E. Lee, "Online usage motive
pp. 109-125, 2010.
and information disclosure for preteen children," Young Consumers, vol. 13, no.
[21] F. Stutzman, R. Capra, and J. Thompson, "Factors mediating disclosure in social
4, pp. 345-356, 2012.
network sites," Computers in Human Behavior, vol. 27, no. 1, pp. 590-598, 2011.
[4] Alkhalifah and J. D’Ambra, "Identity Management Systems Research: Frame-
[22] J. Bryce and J. Fraser, "The role of disclosure of personal information in the
works, Emergemce, and Future Opportunities," in ECIS, 2015.
evaluation of risk and trust in young peoples’ online interactions," Computers in
[5] H. Venter, M. Eloff, L. Labuschagne, J. Eloff, and R. von Solms, New Approaches
Human Behavior, vol. 30, pp. 299-306, 2014.
for Security, Privacy and Trust in Complex Environments: Proceedings of the
[23] F. Annansingh and T. Veli, "An investigation into risks awareness and e-safety
IFIP TC 11 22nd International Information Security Conference (SEC 2007), 14-16
needs of children on the internet: a study of Devon, UK," Interactive Technology
May 2007, Sandton, South Africa. Springer, 2007.
and Smart Education, vol. 13, no. 2, pp. 147-165, 2016.
[6] E. Aghasian, S. Garg, L. Gao, S. Yu, and J. Montgomery, "Scoring users’ privacy
[24] S. Bannon, T. McGlynn, K. McKenzie, and E. Quayle, "The Internet and young
disclosure across multiple online social networks," IEEE Access, vol. 5, pp. 13118-
people with additional support needs (ASN): Risk and safety," Computers in
13130, 2017.
Human Behavior, vol. 53, pp. 495-503, 2015.
[7] K. Li, Z. Lin, and X. Wang, "An empirical analysis of users’ privacy disclosure
[25] J. Bremer, "The internet and children: advantages and disadvantages," Child and
behaviors on social network sites," Information & Management, vol. 52, no. 7, pp.
Adolescent Psychiatric Clinics, vol. 14, no. 3, pp. 405-428, 2005.
882-891, 2015.
[26] J. Wojniak and M. Majorek, "Children in internet space–the European Union
[8] W. Shin and N. Ismail, "Exploring the role of parents and peers in young adoles-
policies on children’s safety online," in SHS Web of Conferences, 2016, vol. 26, p.
cents’ risk taking on social networking sites," Cyberpsychology, Behavior, and
01048: EDP Sciences.
Social Networking, vol. 17, no. 9, pp. 578-583, 2014.
[27] P. Li, T. Li, H. Ye, J. Li, X. Chen, and Y. Xiang, "Privacy-preserving machine
[9] T. Buchanan, C. Paine, A. N. Joinson, and U. D. Reips, "Development of measures
learning with multiple data providers," Future Generation Computer Systems,
of online privacy concern and protection for use on the Internet," Journal of
2018.
the American Society for Information Science and technology, vol. 58, no. 2, pp.
[28] N. Papernot, P. McDaniel, A. Sinha, and M. Wellman, "Towards the science of
157-165, 2007.
security and privacy in machine learning," arXiv preprint arXiv:1611.03814, 2016.
[10] Z. De Souza and G. N. Dick, "Information disclosure on MySpace–the what, the
[29] Z. Ji, Z. C. Lipton, and C. Elkan, "Differential privacy and machine learning: a
why and the implications," Pastoral Care in Education, vol. 26, no. 3, pp. 143-157,
survey and review," arXiv preprint arXiv:1412.7584, 2014.
2008.

123