This chapter explores the negative and positive effects of algorithms on human rights and gender-based discrimination, presents the existing legal framework including its shortcomings and the main constitutive elements of a future human rights law framework on algorithmic gender-based discrimination. Rather than being neutral and overcoming all potential decision-making errors of humans, algorithms perform only as accurately as the design of the algorithm and the underlying (training) data. Data as source of bias, inaccuracy and discrimination can reflect existing inequalities and stereotypes in society and reinforce discrimination. Beyond the regulation of algorithms’ potential negative impacts for society, the chapter argues for an intelligent use of the advantages associated with AI, such as to detect gender discriminations, and more generally achieving less biased and noisy decision making. Although the importance and ubiquitous presence of Artificial Intelligence and the risk of algorithmic discrimination has been recognized in policy documents, no specific legal instrument has been adopted at UN level. Against this background, this chapter sketches out avenues for regulation from a human rights law perspective and it advocates for a hybrid approach ensuring the respect of human rights and the principle of gender equality both at global (UN) and at local (regional and national) level.
https://global.oup.com/academic/product/artificial-intelligence-and-human-rights-9780192882486?cc=ch&lang=en&
https://global.oup.com/academic/covers/pdp/9780192882486