Racial mismatch between applicant and AI avatar increased perceived ethnic bias, while sharing only one identity trait lowered fairness ratings compared to full or no match.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
Adversarial instructions hidden in resumes manipulate LLMs with over 80% attack success, and the proposed FIDS defense with prompt methods reduces attacks by 26.3% while limiting false rejections.
citing papers explorer
-
Skin-Deep Bias: How Avatar Appearances Shape Perceptions of AI Hiring
Racial mismatch between applicant and AI avatar increased perceived ethnic bias, while sharing only one identity trait lowered fairness ratings compared to full or no match.
-
AI Security Beyond Core Domains: Resume Screening as a Case Study of Adversarial Vulnerabilities in Specialized LLM Applications
Adversarial instructions hidden in resumes manipulate LLMs with over 80% attack success, and the proposed FIDS defense with prompt methods reduces attacks by 26.3% while limiting false rejections.