Linguistic Bias in ChatGPT: Language Models Reinforce Dialect Discrimination

TL;DR

This study reveals that ChatGPT models tend to favor standard English dialects and often produce biased, stereotypical, and less accurate responses when interacting with non-standard dialects, highlighting linguistic discrimination issues.

Contribution

The paper provides a large-scale analysis of dialect bias in ChatGPT, demonstrating how models reinforce dialect discrimination and comparing performance between GPT-3.5 and GPT-4.

Findings

Models default to standard English dialects.

Non-standard dialect responses show increased stereotyping and demeaning content.

GPT-4 exhibits more stereotyping despite improved comprehension.

Abstract

We present a large-scale study of linguistic bias exhibited by ChatGPT covering ten dialects of English (Standard American English, Standard British English, and eight widely spoken non-"standard" varieties from around the world). We prompted GPT-3.5 Turbo and GPT-4 with text by native speakers of each variety and analyzed the responses via detailed linguistic feature annotation and native speaker evaluation. We find that the models default to "standard" varieties of English; based on evaluation by native speakers, we also find that model responses to non-"standard" varieties consistently exhibit a range of issues: stereotyping (19% worse than for "standard" varieties), demeaning content (25% worse), lack of comprehension (9% worse), and condescending responses (15% worse). We also find that if these models are asked to imitate the writing style of prompts in non-"standard" varieties, they produce text that exhibits lower comprehension of the input and is especially prone to stereotyping. GPT-4 improves on GPT-3.5 in terms of comprehension, warmth, and friendliness, but also exhibits a marked increase in stereotyping (+18%). The results indicate that GPT-3.5 Turbo and GPT-4 can perpetuate linguistic discrimination toward speakers of non-"standard" varieties.