LLM Targeted Underperformance Disproportionately Impacts Vulnerable Users

TL;DR

This paper investigates how the performance of large language models varies across different user traits, revealing that vulnerable users such as non-native English speakers and those with lower education are disproportionately affected by inaccuracies and undesirable behaviors.

Contribution

The study provides extensive empirical analysis showing that LLMs underperform for vulnerable user groups based on language proficiency, education, and country of origin.

Findings

Undesirable behaviors occur more for users with lower English proficiency.

Lower-education users experience more inaccuracies from LLMs.

Users outside the US face higher rates of model unreliability.

Abstract

While state-of-the-art large language models (LLMs) have shown impressive performance on many tasks, there has been extensive research on undesirable model behavior such as hallucinations and bias. In this work, we investigate how the quality of LLM responses changes in terms of information accuracy, truthfulness, and refusals depending on three user traits: English proficiency, education level, and country of origin. We present extensive experimentation on three state-of-the-art LLMs and two different datasets targeting truthfulness and factuality. Our findings suggest that undesirable behaviors in state-of-the-art LLMs occur disproportionately more for users with lower English proficiency, of lower education status, and originating from outside the US, rendering these models unreliable sources of information towards their most vulnerable users.