WEIRD Populations in LLMs
Abstract
This research examines the extent to which Large Language Models exhibit values and behaviors characteristic of WEIRD (Western, Educated, Industrialized, Rich, and Democratic) populations in economic decision-making contexts.
Description
The WEIRD acronym describes populations that are overrepresented in psychological and economic research. This project investigates whether LLMs, trained primarily on data from these populations, exhibit similar cultural biases in their economic decision-making processes.
Research Question
Large Language Models exhibit values and behaviors characteristic of WEIRD (Western, Educated, Industrialized, Rich, and Democratic) populations in economic decision-making contexts.
Approach
We examine the cultural and ethical biases embedded in LLM responses and decision-making processes, evaluating the extent to which these systems reflect the cultural biases of their training data in economic contexts.
Additional Details
Through cross-cultural economic scenarios and comparative analysis, we assess how LLMs respond to economic choices that may differ across cultural contexts. This research has important implications for the global deployment of AI systems in economic applications.