# Mapping the use of large language models in hiring decisions: a scoping review

## TL;DR

This paper reviews how large language models are being used in hiring decisions and identifies key trends, gaps, and risks in the literature.

## Contribution

A comprehensive scoping review of LLM use in hiring, revealing patterns in application, outcomes, and ethical considerations.

## Key findings

- LLM use in hiring is concentrated in early stages with a focus on efficiency and performance outcomes.
- Ethical risks are widely acknowledged but mitigation strategies lack empirical validation.
- Research is structurally limited in interdisciplinary collaboration and field-based studies.

## Abstract

Large language models (LLMs) are increasingly being explored and deployed across recruitment and selection processes, reshaping how hiring decisions are supported, communicated, and justified. Unlike earlier algorithmic hiring tools, LLMs operate through language-mediated interaction, influencing interpretive and evaluative layers of decision-making. This scoping review maps the academic literature on LLMs in hiring to examine (i) where and how these systems are applied across the hiring pipeline, (ii) what forms of evidence and outcomes are assessed, (iii) which risks and mitigation strategies are documented, and (iv) how disciplinary structures shape research focus. Following PRISMA-ScR guidelines, we synthesize research published between 2018 and 2026 across multiple disciplines using a transparent, lexicon-based coding approach. The results reveal a rapidly expanding but uneven literature, characterized by concentration in early hiring stages, selective outcome measurement favoring efficiency and performance, high awareness of ethical risks with limited empirical validation of controls, and structurally constrained interdisciplinarity. The review highlights key gaps and provides a foundation for future interdisciplinary and field-based research on responsible LLM use in hiring.