For example, you can test a resume filter and find that it isn’t biased against Black people, and it isn’t biased against women. But it might turn out that it is still biased against Black women. This could happen because the filter “fixed” the gender and race bias by over-selecting white women and Black men while under-selecting Black women.
I think this example explains intersectionality in a very clear and practical way. It shows that discrimination is not always simple or obvious. Even if a system seems fair toward different groups separately, it can still unfairly affect people who belong to multiple marginalized groups at the same time. The resume filter example makes the concept easier to understand and shows why fairness in technology is more complicated than people think.