Even AI Thinks It’s a Man’s Job

A PNAS study by Germain Gauthier, Paola Profeta, Debora Nozza and Kenza Elass reveals how AI can reproduce gender inequalities in CV writing and recruitment. Experiments across six language models show that gender stereotypes influence candidate profiles, hiring scores and salary recommendations, particularly in gender-segregated professions. The findings highlight the need to assess algorithmic fairness throughout the recruitment process, as simply instructing AI to be neutral is not enough.

 

Read the full article