AI Hiring Bots Favors Own‑Generated Resumes, Boosting Selection Odds by Up to 60%

Companies increasingly rely on applicant tracking systems (ATS) that scan resumes for keywords and qualifications. Many of these systems now use large language models (LLMs) to evaluate candidates. Job seekers, meanwhile, often turn to the same LLMs to draft or polish their resumes, hoping to beat the automated screen.

Researchers tested AI evaluators against human‑written resumes and LLM‑generated versions. The evaluators were 23% to 60% more likely to select resumes produced by the same LLM they use. In other words, LLMs acting as evaluators favor resumes they themselves generated over those written by humans. The study covered 2,245 real resumes and simulated hiring for 24 occupations, finding the effect strongest in accounting, sales and finance roles.

The self‑preferencing bias can distort hiring outcomes, giving an edge to candidates who match the employer’s AI tool while sidelining equally qualified applicants who do not. Employers may unintentionally overlook strong talent, and job seekers face a new barrier in an already tight market. With tech sector layoffs exceeding 300,000 in the first four months of 2026, the pressure to navigate AI‑driven hiring is intensifying. Watch for regulators and firms to examine audit tools that detect and correct such self‑bias in hiring algorithms.