Evaluating the Usage of African-American Vernacular English in Large Language Models

TL;DR

This paper investigates how well large language models understand and generate African-American Vernacular English (AAVE), revealing significant differences from human usage and highlighting issues of stereotypes and underrepresentation.

Contribution

It provides a comparative analysis of LLMs and human AAVE usage, identifying gaps and biases, and emphasizes the need for more diverse training data and fairness interventions.

Findings

LLMs underuse and misuse AAVE grammatical features

Models replicate stereotypes about African Americans

Significant differences between LLM and human AAVE usage

Abstract

In AI, most evaluations of natural language understanding tasks are conducted in standardized dialects such as Standard American English (SAE). In this work, we investigate how accurately large language models (LLMs) represent African American Vernacular English (AAVE). We analyze three LLMs to compare their usage of AAVE to the usage of humans who natively speak AAVE. We first analyzed interviews from the Corpus of Regional African American Language and TwitterAAE to identify the typical contexts where people use AAVE grammatical features such as ain't. We then prompted the LLMs to produce text in AAVE and compared the model-generated text to human usage patterns. We find that, in many cases, there are substantial differences between AAVE usage in LLMs and humans: LLMs usually underuse and misuse grammatical features characteristic of AAVE. Furthermore, through sentiment analysis and manual inspection, we found that the models replicated stereotypes about African Americans. These results highlight the need for more diversity in training data and the incorporation of fairness methods to mitigate the perpetuation of stereotypes.