Researchers Review AI Evolution from Matching Models to Recruiting Agents

A systematized narrative review published on arXiv explores the evolution of artificial intelligence systems in recruitment. The research documents how the technology has shifted from basic profile matching and behavioral ranking to multi-stage workflows and tool-using recruiting agents.

Based on an analysis of 40 representative works alongside industrial and legal sources, the authors identify three central transitions. These include moving from similarity metrics to reciprocal suitability, transitioning from single models to compound workflows, and shifting evaluation toward evidence- and productivity-aligned standards.

The review also outlines persistent gaps across current systems. According to the findings, privacy is rarely evaluated directly, and existing benchmarks rarely evaluate utility, fairness, privacy, and security in a joint framework.