CONAN - COunter NArratives through Nichesourcing: a Multilingual Dataset of Responses to Fight Online Hate Speech

Abstract

Although there is an unprecedented effort to provide adequate responses in terms of laws and policies to hate content on social media platforms, dealing with hatred online is still a tough problem. Tackling hate speech in the standard way of content deletion or user suspension may be charged with censorship and overblocking. One alternate strategy, that has received little attention so far by the research community, is to actually oppose hate content with counter-narratives (i.e. informed textual responses). In this paper, we describe the creation of the first large-scale, multilingual, expert-based dataset of hate speech/counternarrative pairs. This dataset has been built with the effort of more than 100 operators from three different NGOs that applied their training and expertise to the task. Together with the collected data we also provide additional annotations about expert demographics, hate and response type, and data augmentation through translation and paraphrasing. Finally, we provide initial experiments to assess the quality of our data.

Resources and benchmark corpora for hate speech detection: a systematic review

“…Regarding the languages, as expected, most of the resources use English data, although in some cases they are collected along with texts in Hindi (Bohra et al 2018;Kumar et al 2018a;Mathur et al 2018) or they are part of even larger multilingual collections (Chung et al 2019;Ousidhoum et al 2019;Steinberger et al 2017). It is also worth pointing out that less-resourced languages such as Amharic, Bengali, Slovene and Swedish, are also represented in the corpora we found, thus enabling a greater linguistic diversity in this field.…”

Vindication, virtue, and vitriol

“…They were then extended with minor linguistic variants. This report 15 from Moonshot CVE was used as a guide to the overall conspiracy landscape within COVID-19. They provide some hashtags, and variants were then acquired, again, from looking down the list of hashtags appearing in the dataset for other variants, and including linguistic variations.…”

“…Agonism argues that the contestations of the time can be used to renew democracy and strengthen public discourse [47]. Promising work on recognising an highlighting counter-speech in online communication is already on the horizon [15,50].…”

SWE2: SubWord Enriched and Significant Word Emphasized Framework for Hate Speech Detection

“…We made use of the latter three datasets in our research. Mathew et al [31] and Chung et al [11] provided counter speech datasets for better analysis of hate speech.…”