1. Semantics derived automatically from language corpora necessarily contain human biases;A Caliskan;Phys. Rep.-Rev. Sec. Phys. Lett,2017
2. Gender shades: Intersectional accuracy disparities in commercial gender classification;J Buolamwini;Conference on fairness, accountability and transparency,2018
3. Fairness through awareness;C Dwork;Proceedings of the 3rd innovations in theoretical computer science conference,2012