AI Has Its Own "Favoritism": Study Finds Hiring AIs Tend to Rate Resumes Written by the Same AI Model More Highly

Researchers found that hiring AIs tend to prefer resume summaries generated by the same model over human-written ones.

The bias showed up in comparison tests and in simulated interview-candidate screening, suggesting a self-preference effect.

The study involved models such as GPT-4o, DeepSeek-V3, LLaMA 3.3-70B, and Qwen 2.5-72B.

The finding raises concerns that AI-driven hiring could become less fair if systems favor their own outputs.