Detecting the Hate Code on Social Media

By Rijul Magu, K. Joshi, Jiebo Luo. 2017

87 citations recorded by the source corpus.

Social media has become an indispensable part of the everyday lives of millions of people around the world. It provides a platform for expressing opinions and beliefs, communicated to a massive audience. However, this ease with which people can express themselves has also allowed for the large scale spread of propaganda and hate speech. To prevent violating the abuse policies of social media platforms and also to avoid detection by automatic systems like Google’s Conversation AI, racists have begun to use a code (a movement termed Operation Google). This involves substituting references to communities by benign words that seem out of context, in hate filled posts or Tweets. For example, users have used the words Googles and Bings to represent the African-American and Asian communities, respectively. By generating the list of users who post such content, we move a step forward from classifying tweets by allowing us to study the usage pattern of these concentrated set of users.

Subfields: Artificial Intelligence, Natural Language Processing, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

In Safety, Robustness & Interpretability

Citations in the corpus

HateCheck: Functional Tests for Hate Speech Detection Models (2020, background): “Relatedly, previous work has shown that code words used by hateful users to evade detection are adopted and mostly discarded again within weeks (Magu et al., 2017; Magu and Luo, 2018).”
Recipes for Safety in Open-domain Chatbots (2020, background): “through "dog whistle" communication or coded language) is notably difficult to deal with (Magu et al., 2017; Bhat and Klein, 2020).”
Challenges and frontiers in abusive content detection (2019, background): “Magu and Luo describe how ‘euphemistic’ code words, such as ‘Skype’ or ‘Bing’, are used to derogate particular groups (Magu, Joshi, & Luo, 2017).”
Understanding Abuse: A Typology of Abusive Language Detection Subtasks (2017, background): “(Burnap and Williams, 2015), “most of them come north and are good at just mowing lawns” (Dinakar et al., 2011), “Gas the skypes” (Magu et al., 2017) Table 1: Typology of abusive language. siders whether (i) the abuse is directed at a speciﬁc target, and (ii) the degree to which it is explicit.”
Determining Code Words in Euphemistic Hate Speech Using Word Embedding Networks (2018, methodology): “The movement branched out to a number of social media platforms, and in particular Twitter (Magu et al., 2017).”
Characterizing and Detecting Hateful Users on Twitter (2018, methodology): “Many previous studies collect data by sampling OSNs with the aid of a lexicon with terms associated with hate speech (Davidson et al. 2017; Waseem and Hovy 2016; Burnap and Williams 2016; Magu, Joshi, and Luo 2017).”
Interpretable Multi-Modal Hate Speech Detection (2021, methodology): “We also get the hate code words as explained by Magu et al. (Magu et al., 2017) and obtain those tweets containing such code words along with swear words and identity based derogatory lexicons.”