i10x.ai Unveils Bias in AI-Driven Job Application Evaluations

New Findings Highlight Discrepancies in Resume Assessments by Different AI Models.

AI screens job applications unfairly, depending on which model wrote the resume. A new study by i10X Research reveals a critical flaw in AI-based hiring: the same candidate, with identical qualifications, receives up to 42 percentage points fewer hire recommendations, solely depending on which AI tool wrote the resume.

Catch more HRTech Insights: HRTech Interview With Hari Kolam, CEO and Co-founder of Findem: Featuring Findem’s GliderAI

The analysis is based on 1,576 valid data points across 100 candidate profiles, evaluated by four leading AI systems: GPT-5.4, Claude Sonnet 4.6, Gemini 3 Pro, and Grok 4.3. Key findings: Claude is the strictest evaluator and shows the largest self-bias, hiring only 42% of GPT-written resumes, but 84% of its own.

GPT penalizes its own writing style by 15 percentage points. Gemini-written resumes score highest across all evaluators, averaging a 94.5% hire rate. On one identical document, GPT and Claude diverged by 29 score points, the difference between a borderline maybe and a clear reject.

In automated applicant tracking systems, a “maybe” verdict effectively ends a candidate’s journey. The resume never reaches a human recruiter. This makes the bias measurable, consequential, and urgent to address.

Read More on Hrtech : Why SWIFT is Too Slow for Your Global Workforce?

[To share your insights with us, please write to psen@itechseries.com ]