Order from us for quality, customized work in due time of your choice.
I. Introduction
I.I. Communication & The Internet
It has been almost 30 years since the internet was invented. People of this generation now have access to information and the ability to share information in a way that was never done before. Even though there are arguments against this, online communication adds to the volume of contact to the traditional offline modes of communication through e-mails, chatrooms, instant messaging, and social media.
The sophisticated multi-disciplinary tool enables individuals to connect with people from all over the world. The Internet doesnt have borders, and individuals arent tied down to their geographical location. It brings together topic-based communities. The internet is the largest ungoverned space. There are innumerous pros to this invention and on the other hand, it is also manipulated by cynical forces. People hide behind the screens, there is a lack of accountability, and it provides a platform to spread malicious, hateful, and deceitful content.
I.II. Harassment and hateful content online
In the midst of the social media revolution, hate speech on the internet also grows with it. People experience online harassment on different levels and in different forms. In the milder forms, it harbors negativity and in its severe forms it taints the reputation of an individual, raises privacy concerns, and even poses a threat to physical safety sometimes.
According to the Pew Research Center survey, 41% of Americans have been subjected to some form of online harassment or other. Internet has changed the way people communicate, connecting the world instantaneously but undoubtedly has its negative effects too. The inter-connectedness makes it possible to target individuals, communities, or races and influence opinions or spread hate speech.
I.III. Laws around Hate Speech & Online Harassment
More often than not, hate speech and online harassment fall under free expression. While governments around the world have difficulty passing laws that discourage free speech, European countries have laws in place against hate speech. As baseless conspiracies, polarizing statements, false ideas, and offensive commentary infiltrate our social media, the impending question is whether should there be laws to erase these messages. Should posts like these constitute free expression? Should action be taken by non-governmental entities at the organization level or at the governmental level?
The manifestation of targeted hate messaging is a form of psychological terrorism that exists on a spectrum of severity and those who have experienced severe forms of cyber harassment or cyberbullying have dramatically different reactions and attitudes towards the issue. The Supreme Court has never defined a category of speech as hateful conduct. Hate speech is a part of the First Amendment what it excludes is what is labeled as hate speech. Cyber harassment is not protected under free speech. The harasser(s) can be sued for defamation. Lawsuits require a lot of resources, perpetrators are hard to identify, and, law enforcement will have to employ a considerable amount of forensic expertise to track down individuals who engage in this anonymously.
II. Hate Speech on Social Media
II.I. Proliferation of Hate Speech on Social Media
Social media platforms are an open playfield for online harassment and hate speech. These frequently target ones personal characteristics, appearance, race and ethnicity, and gender. A recent study by Hate Lab shows an increase in hate speech on social media leads to more crimes against minorities in the physical world. The real-world consequences of hate that spreads on social media are another reason why this should be taken seriously.
Language is weaponized and is used as a means to inflict violence. Anonymity is a facilitating factor in the spread of harassment. Online harassment can take many forms from name-calling to targeted campaigns. The people who experience harassment are uncertain whether it was or wasnt harassment. The topic of online harassment and hate speech can be highly subjective on an individuals perception of the two.
II.II. Social Media Community Guidelines
II.II.a. Facebook
Facebook is the place people feel empowered to communicate and they take their role of keeping abuse off their service very seriously. Their guidelines are broken into six different categories. Safety and objectionable content are the clauses that are most important for this project.
Under safety, Facebook acknowledges that bullying and harassment come in different forms, from threats to releasing personally identifiable material, and unwanted malicious content. They believe context and intent matter and have a self-reporting system in place. They also have a bullying prevention hub which is a resource for teenagers, parents, and educators.
Objectionable content covers hate speech and is defined as an attack against people based on characteristics. Hate speech can be shared to raise awareness or for humor but if the intention isnt clear the post is taken down. They separate attacks into three tiers of severity. Apart from monitoring the content on the platform they also have a built-in reporting system in place.
II.II.b. Instagram
Instagram strives to foster a positive and diverse environment by removing content and comments that have possible threats, hate speech, target individuals, degrade them, or shame them. They do allow for stronger conversations to take place. Attacks based on ones race, ethnicity, nationality, sex, gender, or sexual orientation are not OK.
Instagram doesnt allow nudity or photos, videos, or digitally created content that shows sexual intercourse and genitals. Pictures and videos of kids that are nude or partially nude may be taken down due to unanticipated usage by others. Even if it is shared with the right intentions.
Instagram has a built-in report feature that can be used when you see something that violates guidelines. They have a global team that reviews these reports and removes them as quickly as possible. They may remove comments, imagery, or the entire post associated with the comments.
Instagram is an entity of Facebook. Facebook has more extensive guidelines in place and there is no clear mention of whether they apply to Instagram or not.
II.II.c. Twitter
Twitters purpose is to promote public conversation. Violence and harassment in any form diminish the value of public conversation. Twitter’s rules are to ensure all people can participate in the public conversation freely and safely.
II.III. Where do we draw the line?
Social media platforms have played a huge role in social and political protests. From Occupy Wall Street, #MeToo, to BlackLivesMatter it is evident that social media has the potential to create social change. . There is a growing concern about media sexism. While they have guidelines in place these media technologies still suffer from harassment, misogyny, hatred, trolling, and online stalking. This has proliferated to an extent where a recent poll shows that 51% of Americans think that the First Amendment is outdated and should be rewritten and 48% believe hate speech should be illegal.
Free speech shouldnt be used as a shield from the social consequences of your words. Even though a huge percentage of people believe that hate speech should be illegal we should take into account the subjectivity of the matter. So, where do we draw the line? Humans havent been able to distinguish between hate speech and offensive language. They also havent reached a consensus on what constitutes hate speech. The insidious nature of hate speech is that it can take different shapes depending on the context. While there are different opinions on the topic of hate speech itself Facebook has decided to control it using machine learning for its detection.
III. Fighting against online harassment & hate speech
III.I. Machine Learning for Hate Speech Detection
Machines arent like human beings. Their understanding of language is highly mathematical and algorithms can classify text but they are highly sensitive to change. A crucial challenge for machine learning is understanding the context in which it is being communicated. The algorithms are still pre-mature and are susceptible to deception. The system is easy to evade.
Perspective AI is a hate speech machine learning model that is available there and it uses Natural Language Processing to determine the toxicity of a word, sentence, or paragraph. The algorithms of Google perspective view profanity as toxic so when profanity is used in a non-hateful sentence it automatically violates the community guidelines. It also doesnt consider typos and white spaces or the lack thereof. False positives in sentences that provide an aggregate toxic score arent always accurate. These are some of the limitations of the current state of hate speech detection. However, Facebook claims that by using machine learning the rates of removal for hate speech content have increased and that their company removes 72 percent of their illegal hate speech on the platform. I acknowledge by using existing models that the prediction of hate speech might not be fully accurate.
III.II. Project proposal
The project is called Your Enemies Love You and what it aims to do is modify messages of hate and harassment into sentences of self-love and endearment on a social platform. The purpose of the project is to start a bigger conversation about the effects of positive and negative communication amongst communities online. Your enemies love you is a project born out of several explorations during the 7-in-7s.
Project 1: A hate plug-in was a conceptual idea to create a social media credit system based on usage and reporting. If a person has been reported several times and his posts are flagged as harassment or hate speech his credit score is reduced. This would inherently increase accountability online.
Project 2: Minor interventions that prompt before posting to notify the user that a particular message might not adhere to community guidelines. This would also prompt the user to ask if they would want to view a comment on their post or targeted at them that might not adhere to guidelines.
The method used to build will be a web extension. This way users will have the option of choosing to opt into this or not. The project that Im trying to achieve will use a sentiment analysis API to recognize hate speech and harassment and then work towards making individuals rethink before posting and viewing deceitful comments and providing users history of comments that do not adhere to community guidelines.
The project aims at increasing accountability in these ungoverned spaces, promoting ethical speech and expression.
- This is a victim-first approach. The person who is most affected by online harassment and hate speech is the person who is being targeted. Intervention 1 introduces a prompt that alerts the individual before viewing a comment and gives the choice to view the content.
- Think before acting. The second intervention is a prompt that uses hate detection before posting and letting the user know that a comment or a post may violate community guidelines. This allows the user to not post something out of haste or anger and seconds to reconsider before posting something that might be hateful.
- Intervention 3 is what increases the accountability of the user. The intervention aims at using hate speech detection to track the comments and posts of an individual and enlist all the comments that violate the guidelines. It not only increases the accountability of a profile but is also great for the subjectivity of the issue since humans cant agree on hate speech, now they have a list which they can refer to and decide for themselves.
IV. Conclusion
IV.I. Significance of the project
I believe that this project is important and public speech that expresses hate shouldnt be looked at. The growing awareness of the topic and studies that support that hate speech has increased worldwide and can have an impact in real life is reason enough to address online harassment and hate speech. This is a sincere effort to reduce online misconduct. At the end of a computer screen is another human being, it is crucial to see the person behind the screen and not hide behind the anonymity that the Internet offers. This makes social media safer, inclusive, and a little more equitable.
IV.II. Limitations of the project
The chrome extension is not fully built, yet. The project was originally supposed to be built on Twitter due to a lack of technical knowledge and the resilient development of Twitter this had to be foregone. The project aims to use existing hate speech detection models which arent fully accurate at this date.
Bibliography
- Building a Feminist Data Set for a Feminist AI. Akademie Schloss Solitude: Schlosspost, 1 Nov. 2017, https://schloss-post.com/building-feminist-data-set-feminist-ai/.
- Duggan, Maeve. Online Harassment 2017. Pew Research Center: Internet, Science & Tech, Pew Research Center, 4 Jan. 2018, https://www.pewresearch.org/internet/2017/07/11/online-harassment-2017/.
- eschulze9. EU Says Facebook, Google and Twitter Are Getting Faster at Removing Hate Speech Online. CNBC, CNBC, 4 Feb. 2019, https://www.cnbc.com/2019/02/04/facebook-google-and-twitter-are-getting-faster-at-removing-hate-speech-online-eu-finds–.html.
- Matsakis, Louise. To Break a Hate-Speech Algorithm, Try ‘Love’. Wired, Conde Nast, 27 Sept. 2018, https://www.wired.com/story/break-hate-speech-algorithm-try-love/.
- Online Hate Speech Predicts Hate Crimes on the Streets. HateLab, https://hatelab.net/2019/10/14/online-hate-speech-predicts-hate-crimes-on-the-streets/.
- Seaquist, Carla. Free Speech vs. Responsible Speech: We Need to Talk, Again. HuffPost, HuffPost, 5 Apr. 2015, https://www.huffpost.com/entry/free-speech-vs-responsible-speech_b_6563162?guccounter=1&guce_referrer=aHR0cHM6Ly93d3cuZ29vZ2xlLmNvbS8&guce_referrer_sig=AQAAACIPwQIXp53tv37yv02FPUsua3gPyGDtORLoKyL-AXzeUJxy3kOVqSmMSr54mSdqV6iL84XSpjXRs_zpq202GtAC5wWGbk1RAyQl1df3XBqBzet0RP28FTpDhhVzU01zHuxxMTLUwJIW8VV4jZLCWmGyiwb4sXfUzyYc9_Df9nXj.
- The War Against Online Trolls. The New York Times, The New York Times, 3 Dec. 2014, https://www.nytimes.com/roomfordebate/2014/08/19/the-war-against-online-trolls/free-speech-does-not-protect-cyberharassment.
- Gröndahl, Tommi, et al. All You Need Is Love : Evading Hate-Speech Detection. ArXiv:1808.09115 [Cs], Nov. 2018. arXiv.org, http://arxiv.org/abs/1808.09115.
Order from us for quality, customized work in due time of your choice.