Auditing the Use of Language Models to Guide Hiring Decisions
This addresses the need for practical auditing methods to detect algorithmic bias in employment contexts, though it is incremental in adapting existing human bias detection tools to LLMs.
The study tackled the problem of auditing large language models (LLMs) for bias in hiring decisions by applying correspondence experiments to assess race and gender disparities in candidate evaluations for K-12 teaching positions, finding evidence of moderate disparities robust to variations in inputs and task framing.
Regulatory efforts to protect against algorithmic bias have taken on increased urgency with rapid advances in large language models (LLMs), which are machine learning models that can achieve performance rivaling human experts on a wide array of tasks. A key theme of these initiatives is algorithmic "auditing," but current regulations -- as well as the scientific literature -- provide little guidance on how to conduct these assessments. Here we propose and investigate one approach for auditing algorithms: correspondence experiments, a widely applied tool for detecting bias in human judgements. In the employment context, correspondence experiments aim to measure the extent to which race and gender impact decisions by experimentally manipulating elements of submitted application materials that suggest an applicant's demographic traits, such as their listed name. We apply this method to audit candidate assessments produced by several state-of-the-art LLMs, using a novel corpus of applications to K-12 teaching positions in a large public school district. We find evidence of moderate race and gender disparities, a pattern largely robust to varying the types of application material input to the models, as well as the framing of the task to the LLMs. We conclude by discussing some important limitations of correspondence experiments for auditing algorithms.