SELECT LANGUAGE BELOW

AI job screens favor resumes created by AI rather than those written by people, according to researchers.

AI job screens favor resumes created by AI rather than those written by people, according to researchers.

Choose Your Bot Wisely

Job seekers are facing some modern challenges. One of the biggest hurdles now is the applicant tracking system that utilizes AI to prioritize resumes automatically.

A recent study indicates that these AI-powered systems tend to favor resumes generated by AI over those written by humans. Candidates who use the same large-scale language models (LLMs) that companies are using are more likely to get shortlisted.

The study stated that “LLMs systematically prefer self-written resumes over comparable human-written resumes when used as evaluators.”

This bias, identified by researchers Jiangnan Hsu from the University of Maryland, Gujie Li from the National University of Singapore, and Jane Jiang from The Ohio State University, may negatively affect strong candidates. They noted, “If left unchecked, this bias can skew hiring outcomes by systematically favoring candidates who use the same LLM as the employer, while disadvantaging those who do not.” This research was published on arxiv.org.

Emma Wiles, a Boston University professor studying the effects of information systems and AI on the workforce, pointed out that while AI tools are already being used in hiring, the study has revealed new concerns.

“Rather than helping find applicants’ true abilities, these tools are identifying candidates that the AI thinks are similar,” she mentioned.

The researchers analyzed 2,245 human-written resumes and created multiple versions using various LLMs, including the likes of GPT-4o and Deepseek-V3.1. They simulated hiring scenarios for 24 different jobs, discovering that AI evaluators were 23% to 60% more likely to favor candidates using the same LLM.

According to their findings, the issues were particularly pronounced in accounting, sales, and finance roles.

They expressed that such dynamics raise concerns about fairness for job seekers and pose risks for employers who might overlook capable candidates. This bias is perceived as a “new form of bias” that could reshape hiring practices.

Over 300,000 job losses were reported between January and April 2026, especially within the tech industry.

Wiles emphasized that workers should seek tools that enhance human writing skills rather than replace them entirely. “It helps, but it doesn’t replace the hard work you put into writing. It improves on what you’ve already written, it doesn’t shut out your true self,” she said. “Your self is revealed in what you write.”

Facebook
Twitter
LinkedIn
Reddit
Telegram
WhatsApp

Related News