Global-Liar: Factuality of LLMs over Time and Geographic Regions

TL;DR

This study evaluates the factual accuracy, biases, and stability of GPT models across regions and time, introducing a balanced dataset and analyzing how configurations affect reliability, highlighting disparities and areas for improvement.

Contribution

The paper introduces 'Global-Liar,' a novel dataset for evaluating LLM biases across regions and time, and provides insights into how model updates and configurations impact factuality and fairness.

Findings

GPT-4 (March) outperforms later versions in factual accuracy.

Models biased towards Global North, disadvantaging regions like Africa and Middle East.

Binary decision constraints reduce factuality.

Abstract

The increasing reliance on AI-driven solutions, particularly Large Language Models (LLMs) like the GPT series, for information retrieval highlights the critical need for their factuality and fairness, especially amidst the rampant spread of misinformation and disinformation online. Our study evaluates the factual accuracy, stability, and biases in widely adopted GPT models, including GPT-3.5 and GPT-4, contributing to reliability and integrity of AI-mediated information dissemination. We introduce 'Global-Liar,' a dataset uniquely balanced in terms of geographic and temporal representation, facilitating a more nuanced evaluation of LLM biases. Our analysis reveals that newer iterations of GPT models do not always equate to improved performance. Notably, the GPT-4 version from March demonstrates higher factual accuracy than its subsequent June release. Furthermore, a concerning bias is observed, privileging statements from the Global North over the Global South, thus potentially exacerbating existing informational inequities. Regions such as Africa and the Middle East are at a disadvantage, with much lower factual accuracy. The performance fluctuations over time suggest that model updates may not consistently benefit all regions equally. Our study also offers insights into the impact of various LLM configuration settings, such as binary decision forcing, model re-runs and temperature, on model's factuality. Models constrained to binary (true/false) choices exhibit reduced factuality compared to those allowing an 'unclear' option. Single inference at a low temperature setting matches the reliability of majority voting across various configurations. The insights gained highlight the need for culturally diverse and geographically inclusive model training and evaluation. This approach is key to achieving global equity in technology, distributing AI benefits fairly worldwide.