Intersectionality in AI: Challenges and Solutions for Multidimensional Equity
The integration of AI into various sectors (recruitment, banking, hospitals, the judicial system) raises crucial questions regarding equity and the fight against discrimination. The reality of discriminatory practices necessitates an “intersectional” approach, which takes into account the combinations of identity characteristics—such as gender or age—in individuals’ profiles, rather than their isolated effects. Addressing bias from this perspective complicates the calculation of a model’s bias level, particularly due to the resulting multidimensionality. This approach is, however, necessary to ensure the responsible use of AI.
Read the article