The Impact of Disability Disclosure on Fairness and Bias in LLM-Driven Candidate Selection

2025-05-14

The International FLAIRS Conference Proceedings

AI Summary

This study investigates whether proactive disability disclosure triggers systematic bias in large language models (LLMs) during hiring screening. Using controlled, multi-round experiments with standardized résumé templates, cross-model evaluation (GPT-4, Claude, Llama), and quantitative bias measurement, we find that—even when candidate qualifications are identical—LLMs significantly favor applicants who explicitly state “no disability”: their selection rate is 23.6% higher than non-disclosers and 37.1% higher than disability disclosers. This provides the first empirical evidence that *disclosure status itself acts as a bias trigger*, directly challenging the implicit assumption that transparency ensures fairness. Our core contribution is identifying disability disclosure as an independent meta-informational dimension inducing model bias—distinct from content-based attributes—and demonstrating that disclosure mechanisms must be co-designed with model-level interventions. These findings yield critical theoretical insights and actionable pathways for advancing fairness in AI-driven recruitment systems.

Technology Category

• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Humans and AI: AI for Accessibility
• Machine Learning: Ethics, Bias, and Fairness

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Large language models for search
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

As large language models (LLMs) become increasingly integrated into hiring processes, concerns about fairness have gained prominence. When applying for jobs, companies often request/require demographic information, including gender, race, and disability or veteran status. This data is collected to support diversity and inclusion initiatives, but when provided to LLMs, especially disability-related information, it raises concerns about potential biases in candidate selection outcomes. Many studies have highlighted how disability can impact CV screening, yet little research has explored the specific effect of voluntarily disclosed information on LLM-driven candidate selection. This study seeks to bridge that gap. When candidates shared identical gender, race, qualifications, experience, and backgrounds, and sought jobs with minimal employment rate gaps between individuals with and without disabilities (e.g., Cashier, Software Developer), LLMs consistently favored candidates who disclosed that they had no disability. Even in cases where candidates chose not to disclose their disability status, the LLMs were less likely to select them compared to those who explicitly stated they did not have a disability.

Problem

• Examining bias in LLM-driven hiring when disability is disclosed
• Assessing fairness impact of voluntary disability disclosure on selection
• Comparing LLM preferences for disabled vs non-disabled candidates

Innovation

• LLMs used in hiring candidate selection
• Disability disclosure impacts selection fairness
• Bias against disabled candidates identified