Recruitment agency Oriel Partners collected 100 “real” CVs for a specific role in its own firm and used ChatGPT to enhance them in relation to the job description. It then compared the AI-enhanced CVs with the original ones to see how they scored and to gauge the extent of modifications.
It found that ChatGPT lied on average 14 times on CVs. These embellishments included additions to the profile section, improving applicants’ key skills and attributes, and enhancing their professional experience.
The AI-enhanced CVs then went on to score significantly higher when a custom ChatGPT-powered tool was deployed to screen the CVs, suggesting that candidates using AI to improve their selection chances could be placing themselves at an unfair advantage.
