LLM Targeted Underperformance Disproportionately Impacts Vulnerable Users
This work highlights a fairness issue in AI where LLMs underperform for vulnerable users, potentially exacerbating inequalities in access to reliable information.
The study investigated how the quality of LLM responses varies based on user traits like English proficiency, education level, and country of origin, finding that undesirable behaviors such as inaccuracies and refusals occur disproportionately more for vulnerable users, such as those with lower English proficiency or from outside the US.
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.