Evaluation of LLMs Biases Towards Elite Universities: A Persona-Based Exploration

TL;DR

This paper examines biases in popular LLMs towards elite universities when generating professional personas, revealing significant overrepresentation of these institutions compared to real-world data, and discusses implications for AI bias mitigation.

Contribution

Introduces a novel persona-based method to quantify educational biases in LLMs and compares multiple models against real LinkedIn data.

Findings

LLMs overrepresent elite universities 72.45% of the time.

ChatGPT 3.5 shows the highest bias.

Gemini performs better with less bias.

Abstract

This study investigates whether popular LLMs exhibit bias towards elite universities when generating personas for technology industry professionals. We employed a novel persona-based approach to compare the educational background predictions of GPT-3.5, Gemini, and Claude 3 Sonnet with actual data from LinkedIn. The study focused on various roles at Microsoft, Meta, and Google, including VP Product, Director of Engineering, and Software Engineer. We generated 432 personas across the three LLMs and analyzed the frequency of elite universities (Stanford, MIT, UC Berkeley, and Harvard) in these personas compared to LinkedIn data. Results showed that LLMs significantly overrepresented elite universities, featuring these universities 72.45% of the time, compared to only 8.56% in the actual LinkedIn data. ChatGPT 3.5 exhibited the highest bias, followed by Claude Sonnet 3, while Gemini performed best. This research highlights the need to address educational bias in LLMs and suggests strategies for mitigating such biases in AI-driven recruitment processes.