English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Language is a dynamic aspect of our culture that changes when expressed indifferent technologies/communities. Online social networks have enabled thediffusion and evolution of different dialects, including African AmericanEnglish (AAE). However, this increased usage is not without barriers. Oneparticular barrier is how sentiment (Vader, TextBlob, and Flair) and toxicity(Google's Perspective and the open-source Detoxify) methods present biasestowards utterances with AAE expressions. Consider Google's Perspective tounderstand bias. Here, an utterance such as ``All n*ggers deserve to dierespectfully. The police murder us.'' it reaches a higher toxicity than``African-Americans deserve to die respectfully. The police murder us.''. Thisscore difference likely arises because the tool cannot understand there-appropriation of the term ``n*gger''. One explanation for this bias is thatAI models are trained on limited datasets, and using such a term in trainingdata is more likely to appear in a toxic utterance. While this may beplausible, the tool will make mistakes regardless. Here, we study bias on twoWeb-based (YouTube and Twitter) datasets and two spoken English datasets. Ouranalysis shows how most models present biases towards AAE in most settings. Weisolate the impact of AAE expression usage via linguistic control features fromthe Linguistic Inquiry and Word Count (LIWC) software, grammatical controlfeatures extracted via Part-of-Speech (PoS) tagging from Natural LanguageProcessing (NLP) models, and the semantic of utterances by comparing sentenceembeddings from recent language models. We present consistent results on how aheavy usage of AAE expressions may cause the speaker to be consideredsubstantially more toxic, even when speaking about nearly the same subject. Ourstudy complements similar analyses focusing on small datasets and/or one methodonly.

A Comprehensive View of the Biases of Toxicity and Sentiment Analysis Methods Towards Utterances with African American English Expressions