Literature Review: Hate Speech on Twitter
Table of Contents
- Literature Search2
1.1 Introduction2
1.2 Keywords2
1.3 Datasets3
- Literature Review8
2.1 Research Statement11
- Scope and Limitations of the Research12
- Conclusion13
- References14
Literature Search
1.1 Introduction
With the popularity and transformation of social media as a powerful platform to share and exchange ideas, people around the world have also taken to posting hate speeches and comments on such platforms. One such social media channel that has been in the limelight for hatred speeches and comments by users in recent years is Twitter. Twitter has become a channel for those involved with racial slurs, derogatory tweets on the basis of racial or religious discrimination, and to display sentiments of sexism, xenophobia, and homophobia against other users on the Internet. This trend of using hate remarks and speeches to instigate and violate the privacy, rights, and sentiments of others has become a concern for online users, authorities and government institutions around the world (Chen et al. 2012). Thus, the research paper focuses on classifying the degree of Twitter hate speeches in different parts of the world. The research will include the demographics of users involved with sexist, racial, religiously discriminative, xenophobic, and homophobic comments and speeches on Twitter. The datasets will be followed by a literature review that identifies the reasons behind hate speeches on social media (particularly Twitter) and assesses its impact on the society as a whole.
1.2 Keywords
Twitter hate speeches, sexist speeches, sexism, xenophobia on Twitter, homophobic comments on Twitter, racial slurs, religious discrimination, religious offense
1.3 Datasets
Figure 1: Representation of professions
Figure 2: Representation of User Interests
Figure 3: Xenophobia by Twitter users based on Gender
Figure 4: Representation of Homophobic sentiments based on % of Twitter mentions
Figure 5: Representation of amount of Hate speech used in top 3 countries
Figure 7: Representation of five kinds of hate speech
Figure 7: Representation of responses to Hate speech by Twitter Users
Figure 8: Representation of Twitter Hate speech on Sexism based on Gender.
Literature Review
The use of hate speeches and hateful remarks/comments on social media websites has been proliferating at a rapid rate on social media platforms. In ‘Understanding Harmful Speech Online’ (2017), keen observations and reflections are provided into appropriating strategies and measures that can further assess the impact of online hate speeches better. As Twitter is the most popular platform for analyzing the different kinds of hate speeches used by people from around the world, a detailed review has been provided highlighting deeper insights into how social, civil, and sentiment challenges from hate speeches can be mitigated and the role of research, civil bodies, government institutions, and policymakers of balancing the impact of such speeches on society. The study suggests that not all kinds of hateful remarks, comments, or tags on Twitter can be monitored with equilibrium. Thus, the focus needs to shift towards analyzing the impact of hate speeches based on the classification of the degree they fall into (Allcott & Gentzko 2017). On one hand, where such public interventions on social media are considered as a radical form of expressions that asserts freedom to online users, on the other hand, extremely threatening comments and speeches made by Twitter users can have a drastic impact on the psychological, sociological, and sentimental factors of the society. Additionally, the paper conducts an analysis of the prevalence and significance of Twitter hate speech that occurs on the lines of religious/racial discrimination, xenophobia, homophobia, and primarily, sexism. Attempts have been made by the gatekeepers, policymakers, and platforms like Twitter to track and classify the different types of hate speech used online. However, there is a thin line of difference between what can be categorized as vague, negative, offensive, and extreme (Twitter team 2017). This calls for more detailed and comprehensive research to categorize the degree and extent of Twitter hate speeches so as to analyze and critically evaluate their positive and negative impacts on society. Owing to the lack of evidence on potentially dangerous hate speeches, racial slurs, offensive speeches/words, etc., has led to greater challenges in the way derogatory remarks and discriminating speeches are handled on social media like Twitter. Hence, the authors propose different research strategies and shed light on the roles of community, policymakers, legal institutions, and intermediaries to provide more reflection, which can help understand the difference and varying impacts of threatening and non-threatening Twitter hate speech. The paper concludes by suggesting that there is still a lot of scope and progress to be achieved when it comes to classifying online hate speech and counterspeech that can necessitate the need for social, civil, and legal interventions (Jha & Mamidi 2017).
‘What Is Hate Speech? Part 1: The Myth Of Hate’ by Alexander Brown (2017) defines hate speech as a broader time than just using incriminating, offensive, defaming, insulting, stereotyping, or degrading remarks. There are four ways in which the study differentiates and goes beyond the legal definition and concept of hate speech. First, hate speech constitutes for more general remarks and comments that are directed towards religion, ethnicity, culture, gender identity, nationality, race, sexual orientation (e.g. xenophobia and homophobia), and disability (Wrench 2016). It has been identified that people involved in using hate speech on social media platforms like Twitter take up to such acts as a part of their democratic right and freedom; to counterattack and argue in order to put across their ideas, views, and beliefs. However, the increasing and rampant use of online hate speech are rather destructive in nature that further hampers the way democracy and freedom of expression is exercised, endangering that very right of the society (Brown 2017). Second is to lay emphasis on the characteristics that identify the groups of the population involved in Twitter hate speech. This will help in better and more accurate group classification that can be associated with the legal concept of hate speech. The nature of speech is the third aspect that the study highlights, suggesting that it is important to differentiate between what is acceptable and what is not; for instance, assessing characteristics of its nature as an offensive speech or insulting speech on the basis of the words used and the nature of the speech. The fourth suggestion is to break up the concepts of hate speech further in order to identify the sub roles that insulting and offensive speeches can take. This includes identifying specific kinds of speech directed towards terrorizing, discrimination, defaming, humiliating, persecution, and more (Sunstein 2018). As such, the study analyses the perspective of hate speech users and the effects of different kinds of hate speeches, with the main focus on hate speech made on the basis of race, religion, sexism, xenophobia, and homophobia in this research paper. Additionally, this specific review will provide assistance in evaluating the degree of Twitter hate speech into non-threatening, threatening, and extremely-threatening (Liu 2012).
‘A Measurement Study of Hate Speech in Social Media’ (Mondal, Silva, & Benevenuto 2017) is a comprehensive research that attempts to identify the patterns of hate speech in different countries. The authors provide a broader perspective on the meaning of hate speech and how different techniques can be used to understand and analyze the multifaceted nature of online hate speech. The research involves both qualitative and quantitative data collection from social media websites. Data collected from Twitter highlight the different kinds of hate speeches from users belonging from different parts of the world. Additionally, the issues pertaining to hate speech and anonymity have also been discussed and variations explained from one country to another. According to the geographical detection of Twitter hate speech, it was found that the hatred or hateful sentiments were much wider between countries, with little variation of sentiments of hatred in different cities within a country (Ribeiro et al. 2017). Including Wordtree visualization, the study explains the context of online hate speech on the basis of the nature of words and sentences used. The implications of this selected literature can be seen in mitigating both online and offline sexual discrimination and bias prevalent in countries like the United Kingdom and the United States, primarily on the basis on xenophobic and homophobic sentiments. In addition, the relation between hate speech/hate users and anonymity must be clearly identified and justified in order to establish stronger identities of individuals across social media platforms like Twitter. New hate targets must be evaluated and monitored to ensure that keywords used to detect hate speech on Twitter can be updated and more relevant to the kind of words being used by haters online. It also indicates the need for more real-time keyword detection software that can identify the kind and impact of hate speech posted on Twitter in the shortest possible time, before it ends up having a negative influence on online users or society (Waseem & Hovy 2016). Considering the tremendous effort, resource and time being spent into monitoring and curbing the use of hate speech on Twitter, the authors in this research lay important on understanding and analyzing the vastness of online hate speech and the abundant number of online hate users across different geographical regions. It is required that hate speech is classified and identified more precisely, instead of having to curb freedom of speech and expression to a drastic level. By identifying the degree, category, and pattern of hate speech being used, necessary and ethical preventive measures can be taken to reduce negative influences generated by hate users and to transit to a positive aspect of how public interventions can be beneficial to the society.
2.1 Research Statement
To understand the broader context of ‘hate speech’ and classify Twitter hate speeches according to their threatening and non-threatening natures.
To evaluate the different kinds of hate speech posted by users across different geographical locations and demographics.
To encourage open-ended discussions and research on the positive and negative influences of hate speech and their lasting impact on the society.
Scope and Limitations of the Research
This research has captured the broader context of hate speech used online, specifically on Twitter. Hate speech affects people in different ways, and hence, it is important to understand and evaluate the nature of different kinds of hate speech to draw a line between views and beliefs that are threatening and non-threatening for people and the society as a whole. Due to the abundance of hate speeches used online and the complexity of its meaning and context has immense scope for research, so as to introduce ethical measures that can respect freedom of expression, while assessing the sentiment analysis and negative social influence caused by such remarks (Faris, Ashar & Gasser 2016). Not all kinds of hate speech can be monitored on the basis of certain keywords, on which innovative machine learning techniques are being developed to reduce hateful users from hurting others on social media platforms like Twitter. Thus, further research and identification of the right categories and context of online hate speech can help bridge its gap with anonymity, geography, and other demographical factors (Khoza 2017).
The limitations of the research are the non-inclusion of cities as the project is limited to analyzing hate speech between countries, and not cities within countries. Another probable limitation of the paper is the lack of research found on the age groups of Twitter users who are involved in hate speech and regularly post hateful comments and hashtags. Additionally, this study is inclined towards evaluating hate speech on the basis of five determinants; (1) sexism, (2) racism, (3) religious discrimination, (4) xenophobia, and (5) homophobia. However, hate speech comprises of a wider range of classifications on the basis of behaviors, class, identity,
Conclusion
This research paper aims at identifying the differences between threatening, non-threatening and extremely-threatening hate speech on Twitter. An analysis of different kinds of online hate speech provides a deeper understanding of the broader context and meaning of using hate speeches, especially on social media platforms like Twitter. By classifying threats and violent speech posted by users to express anger and hatred towards others can help improve the way in which hate speech is monitored across different social media channels. Additionally, maintaining ethics by identifying the fine line of difference between democracy and freedom of speech and expressions can help reflect and observe the topic of online hate speech, which can motivate researchers to precisely evaluate the positive and negative influences of different online attitudes on society.
References
Allcott, H & Gentzko, M 2017, “Social media and fake news in the 2016 election”, Journal of Economic Perspectives, vol. 31, no. 2, pp. 211-36.
Assimakopoulos, Stavros & Rebecca Vella Muskat. 2017. Exploring xenophobic and homophobic attitudes in Malta through the analysis of online comments to news stories. Lodz Papers in Pragmatics 13(2). 179–202.
Bartlett, J & Richard, N 2015, Immigration on Twitter: understanding public attitudes online, viewed on 5 April 2019, Retrieved from: http://apo.org.au/node/54788
Brown, A 2017, “What is hate speech? Part 1: The Myth of Hate”, Law and Philosophy, vol. 36, no. 4, pp. 419-468.
Chen, Y, Zhou, Y, Zhu, S, & Xu, H 2012, “Detecting offensive language in social media to protect adolescent online safety”, 2012 International Conference on Privacy, Security, Risk and Trust (PASSAT 2012), pp. 71–80.
Elanor Colleoni, Alessandro Rozza and Adam Arvidsson. 2014. Echo chamber or public sphere? Predicting political orientation and measuring political homophily in Twitter using big data. Journal of Communication, 64(2): 317-332
Faris, R., Ashar, A. and Gasser, U. (2016). Understanding Harmful Speech Online. SSRN Electronic Journal, no. 21.
Jha, A & Mamidi, R 2017, “When does a compliment become sexist? analysis and classification of ambivalent sexism using twitter data”, Proceedings of the Second Workshop on NLP and Computational Social Science, pp. 7–16.
Khoza, A 2017, New social media research finds xenophobia rife among South Africans. news24. viewed on 5 April 2019, Retrieved from http://www.news24.com/SouthAfrica/News/new-social-media-research-finds-xenophobia-rife-among-south-africans-20170404
Liu, B 2012, “Sentiment analysis and opinion mining. Synthesis Lectures on Human Language Technologies, vol. 5, no.1, pp. 1-167
Matthew C. Benigni, Kenneth Joseph and Kathleen M. Carley. 2017. Online extremism and the communities that sustain it: Detecting the ISIS supporting community on Twitter. PLOS One, 12.
Mondal, M, Silva, LA, & Benevenuto, F 2017, “A Measurement Study of Hate Speech in Social Media”, Proceedings of the 28th ACM Conference on Hypertext and Social Media, pp. 85-94
Ribeiro, MH, Calais, P, Santos, Y, Almeida, V & Meira, W 2017,”Like sheep among wolves’: Characterizing hateful users on Twitter”, 11th ACM International Conference on Web Search and Data Mining.
Sunstein, CR 2018, #Republic: Divided democracy in the age of social media, Princeton University Press.
Twitter team 2017, The Streaming APIs, viewed on 5 April 2019, Retrieved from https://dev.twitter.com/streaming/overview.
Waseem, Z & Hovy, D 2016, “Hateful symbols or hateful people? Predictive features forhate speech detection on Twitter”, Proceedings of the North American Chapter of the Associationfor Computational Linguistics (NAACL) Student Re-search Workshop, pp. 88–93,
Wrench, J 2016, Diversity management and discrimination: Immigrants and ethnic minorities in the EU, Routledge: New York.