Despite increasing media attention on biases and stereotypes in databases used by algorithms, reporting on frequent cases of algorithmic gender-based discriminations in all spheres of life and slowly emerging discussions in international policy circles, specific legal rules on gender equality addressing the problems and opportunities of AI for women and the concept of algorithmic discrimination are a rarity. However, the 1979 CEDAW text is drafted in an open and forward-looking way, does not only address discrimination ‘in all its forms’ but also calls upon states parties to address discrimination ‘by all appropriate means’ which notably includes ‘appropriate legislation’ (Art. 2). The substantive provisions of CEDAW could therefore be seen as an invitation for State parties to adopt legislation and for the CEDAW Committee to interpret and embrace the concept of algorithmic discrimination in its Recommendations and Observations.
‘Decision-making systems’ should be understood notably to include AI systems and algorithms that are increasingly used in public and private decision-making, either replacing or complementing human decision-making.
It is argued that the 40th General Recommendation should not only consider and take note of the different positive and negative impacts of AI and algorithms on gender equality and women’s rights but pave the way for a future legal and political framework to ensure that the core principles of CEDAW are equally respected in the algorithmic age. The Recommendation should give concrete guidance on how CEDAW can be interpreted in light of the increased usage of AI and its impacts on women and girls, notably in relation to algorithmic discrimination.