An official website of the United States government
Large language models (LLMs) are fast becoming ubiquitous and have shown impressive performance in various natural language processing (NLP) tasks. Annotating data for downstream applications is a resource-intensive task in NLP. Recently, the use of LLMs as a cost-effective data annotator for annotating data used to train other models or as an assistive tool has been explored. Yet, little is known regarding the societal implications of using LLMs for data annotation. In this work, focusing on hate speech detection, we investigate how using LLMs such as GPT-4 and Llama-3 for hate speech detection can lead to different performances for different text dialects and racial bias in online hate detection classifiers. We used LLMs to predict hate speech in seven hate speech datasets and trained classifiers on the LLM annotations of each dataset. Using tweets written in African-American English (AAE) and Standard American English (SAE), we show that classifiers trained on LLM annotations assign tweets written in AAE to negative classes (e.g., hate, offensive, abuse, racism, etc.) at a higher rate than tweets written in SAE and that the classifiers have a higher false positive rate towards AAE tweets. We explore the effect of incorporating dialect priming in the prompting techniques used in prediction, showing that introducing dialect increases the rate at which AAE tweets are assigned to negative classes.
more »
« less
The Risk of Racial Bias in Hate Speech Detection
We investigate how annotators’ insensitivity to differences in dialect can lead to racial bias in automatic hate speech detection models, potentially amplifying harm against minority populations. We first uncover unexpected correlations between surface markers of African American English (AAE) and ratings of toxicity in several widely used hate speech datasets. Then, we show that models trained on these corpora acquire and propagate these biases, such that AAE tweets and tweets by self-identified African Americans are up to two times more likely to be labelled as offensive compared to others. Finally, we propose dialect and race priming as ways to reduce the racial bias in annotation, showing that when annotators are made explicitly aware of an AAE tweet’s dialect they are significantly less likely to label the tweet as offensive.
more »
« less
- Award ID(s):
- 1714566
- PAR ID:
- 10107991
- Date Published:
- Journal Name:
- ACL
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
ISCA (Ed.)In this paper, we explore automatic prediction of dialect density of the African American English (AAE) dialect, where dialect density is defined as the percentage of words in an utterance that contain characteristics of the non-standard dialect. We investigate several acoustic and language modeling features, including the commonly used X-vector representation and ComParE feature set, in addition to information extracted from ASR transcripts of the audio files and prosodic information. To address issues of limited labeled data, we use a weakly supervised model to project prosodic and X-vector features into low-dimensional task-relevant representations. An XGBoost model is then used to predict the speaker's dialect density from these features and show which are most significant during inference. We evaluate the utility of these features both alone and in combination for the given task. This work, which does not rely on hand-labeled transcripts, is performed on audio segments from the CORAAL database. We show a significant correlation between our predicted and ground truth dialect density measures for AAE speech in this database and propose this work as a tool for explaining and mitigating bias in speech technology.more » « less
-
Hate speech and offensive language are rampant on social media. Machine learning has provided a way to moderate foul language at scale. However, much of the current research focuses on overall performance. Models may perform poorly on text written in a minority dialectal language. For instance, a hate speech classifier may produce more false positives on tweets written in African-American Vernacular English (AAVE). To measure these problems, we need text written in both AAVE and Standard American English (SAE). Unfortunately, it is challenging to curate data for all linguistic styles in a timely manner—especially when we are constrained to specific problems, social media platforms, or by limited resources. In this paper, we answer the question, “How can we evaluate the performance of classifiers across minority dialectal languages when they are not present within a particular dataset?” Specifically, we propose an automated fairness fuzzing tool called FuzzE to quantify the fairness of text classifiers applied to AAVE text using a dataset that only contains text written in SAE. Overall, we find that the fairness estimates returned by our technique moderately correlates with the use of real ground-truth AAVE text. Warning: Offensive language is displayed in this manuscript.more » « less
-
Abstract Social stereotypes negatively impact individuals’ judgments about different groups and may have a critical role in understanding language directed toward marginalized groups. Here, we assess the role of social stereotypes in the automated detection of hate speech in the English language by examining the impact of social stereotypes on annotation behaviors, annotated datasets, and hate speech classifiers. Specifically, we first investigate the impact of novice annotators’ stereotypes on their hate-speech-annotation behavior. Then, we examine the effect of normative stereotypes in language on the aggregated annotators’ judgments in a large annotated corpus. Finally, we demonstrate how normative stereotypes embedded in language resources are associated with systematic prediction errors in a hate-speech classifier. The results demonstrate that hate-speech classifiers reflect social stereotypes against marginalized groups, which can perpetuate social inequalities when propagated at scale. This framework, combining social-psychological and computational-linguistic methods, provides insights into sources of bias in hate-speech moderation, informing ongoing debates regarding machine learning fairness.more » « less
-
Understanding How Dialect Differences Shape How AAE-Speaking Children Process Sentences in Real-TimeAbstract Dialect differences between African American English (AAE) and Mainstream American English (MAE) impact how children comprehend sentences. However, research on real-time sentence processing has the potential to reveal the underlying causes of these differences. This study used eye tracking, which measures how children interpret linguistic features as a sentence unfolds, and examined how AAE- and MAE-speaking children processed “was” and “were,” a morphology feature produced differently in MAE and AAE. Fifty-nine participants, ages 7;8 to 11;0 years, completed standardized measures of dialect density and receptive vocabulary. In the eye tracking task, participants heard sentences in MAE with either unambiguous (e.g., “Jeremiah”) or ambiguous (e.g., “Carolyn May”), subjects and eye movements were measured to singular (image of one person) or plural referents (image of two people). After the onset of the auxiliary verb, AAE-speaking children were sensitive to “was” and “were” when processing sentences but were less likely than MAE-speaking children to use “was” as a basis for updating initial predictions of plural referents. Among African American children, dialect density was predictive of sensitivity to “was” when processing sentences. Results suggest that linguistic mismatch impacts how contrastive verb morphology is used to update initial interpretations of MAE sentences.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
