IEEE Technology and Society Magazine - September 2020 - 56

these profiles were posting similar tweets; although
their content was not malicious. Moreover, these
profiles were created at nearly the same point in
time and later suspended by Twitter.
■■ We also performed topical analysis of trapped users
content to glimpse their topical inclination. Figure 5
shows top-ten topics and depicts that the most relevant and discussed topics, such as finance, computer accessories, and shopping are advertising-related
topics. They also discussed important topics like
education and society, but relatively less frequently.

Threat Dimensions
In the early stage of research, socialbots were developed and deployed in OSNs to observe their infiltration
potential and do impact analyses. As an open-source
initiative, Coburn and Marra [9] conducted a socialbots
injection experiment on Twitter under the Realboy project. In the project, socialbots imitated the sense-thinkact 5 model of physical robots to obser ve the
information propagation behavior of social robots
(socialbots) and underlying impact on the social graph.
This was an early study, and since then researchers
have repeatedly reported misuse cases in different contexts of the OSNs [1]-[3]. Mitter et al. [10] presented a
generic framework to categorize socialbot attacks
based on five parameters - target, account type, vulnerability, attack method, and result. However, authors
have not categorized socialbots based on their intention

2,000
1,500
1,000
500

Pets

Travel

Style and Fashion

Education

Shopping

Society

Finance

Health and Fitness

Art and Entertainment

0
Technology and Computing

Discussion Frequency

2,500

Topics

FIGURE 5. Top-10 topics discussed by the socialbot followers.
5

A robotic paradigm which describes the three primitives of a robot.
"Sense" represents data gathering (e.g., motion detection) process of
robots; "think" represents information processing power; and finally "act"
represents the decision making ability.

and functionality. In this study, we have identified different threat dimensions where socialbots abuse OSN platforms for deceptive activities.

Socialbots as Political Tools
OSNs are modern discussion platforms and open in
nature where people debate government policies, daily
life events, social issues, and so on. In this era, users
purchasing behavior, political inclinations, etc., are
influenced by social media events and discussion [11],
[12]. Business organizations and political parties are
using social media platforms for targeted advertisements, to log and resolve customer grievances, to
observe users' opinions regarding policies and products, etc. The huge user base, open nature, and easy-touse functionalities of OSNs have also attracted
adversaries. Socialbot handlers exploit the open nature
of OSNs at the top of social engineering to carry out
smear campaigns against political rivals [13], [14]. The
first misuse of socialbots as a political tool was reported in the 2010 U.S. midterm election [15]. Since OSNs
are generally used to express views regarding current
incidents, political scenarios, and daily life events, carrying out smear campaigns will hamper the ultimate
objectives and growth of the underlying OSN platforms.
Political parties around the world are using socialbots to
distort public opinion and create artificially manufactured campaigns [14]. Dickerson et al. [13] also reported the use of socialbots in the 2014 Indian general
election and presented a sentiment and linguistic features-based technique for socialbots detection. The
existing studies report the use of these computational
agents in most of the larger democracies. The problem
is not just the use of socialbots but their impact on
results. Moreover, government authorities inject socialbots to bias public opinion for pro-regime discussions
[1]. Socialbots as a political tool are used for astroturfing, fame fostering, opinion manipulation, etc., as
shown in Figure 6.

Socialbots as Fake News Distributors
The fake news problem is not new and dates back to the
printing press era [16]. In traditional news media, scope
and repercussions of fake news are limited, and its
aftereffects can be controlled and sometimes reversed.
Traditional news media does not have the magnitude of
the user base and real-time proliferation power like
OSNs, which have provided an unprecedented opportunity to use antisocial elements to misuse these platforms for rumor and fake news propagation. Malicious
users exploit socialbots for fake news and misinformation diffusion to pollute social media discourse. Social
media are modern news and information sharing platform; therefore, ensuring the credibility and authenticity

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

∕

SEPTEMBER 2020

IEEE Technology and Society Magazine - September 2020

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - September 2020