'Self-Preference' | The New Recruitment Bias: How AI Hiring Tools May Favour AI-Written CVs

Why It Matters

If AI systems preferentially rank AI‑crafted CVs, job seekers without access to such tools face systemic disadvantage, raising legal and ethical challenges for employers.

Key Takeaways

• AI screening tools may unintentionally reward AI‑crafted resumes
• Human oversight diminishes as LLMs automate candidate ranking
• Detection of AI‑generated content becomes unreliable without human review
• Bias could exacerbate inequality for candidates lacking AI resources

Pulse Analysis

Employers have long worried that candidates will lean on generative AI to polish applications, fearing a loss of authenticity. Yet the same technology is now being deployed on the employer side, with large language models parsing thousands of resumes, scoring keywords, and even drafting interview questions. This automation promises speed and consistency, but it also removes the human intuition that once caught overly polished language or subtle inconsistencies, reshaping the recruiter’s role into a largely algorithmic gatekeeper.

Recent academic studies reveal a paradox: AI‑powered screening tools may inadvertently reward the very AI‑generated content they were meant to flag. Because these systems are trained on vast corpora of high‑quality text, resumes crafted with LLMs often align more closely with the models’ internal language patterns, leading to higher relevance scores. The bias is subtle—candidates who use AI tools gain a hidden advantage, while those who write manually may be penalized despite comparable qualifications. This dynamic challenges traditional fairness metrics and could expose firms to discrimination claims if the bias disproportionately affects under‑resourced applicants.

For businesses, the emerging bias signals a need for transparent AI governance. Companies should audit their screening algorithms for preferential treatment of AI‑written text, incorporate human checks at critical decision points, and consider offering AI assistance uniformly to all applicants to level the playing field. Regulators are beginning to scrutinize algorithmic fairness in hiring, and proactive steps now can mitigate legal risk while preserving the efficiency gains AI promises. Ultimately, balancing automation with equitable outcomes will define the next wave of responsible recruitment technology.