PROBLEM FRAMING

Discrimination against statutorily protected classes has been increasingly documented in online environments. Concerns include how the design of online platforms, the pervasive collection of data, automated decisions, and scant control over users’ activities might enhance forms of structural discrimination, such as the denial of ethnic minorities from the access of goods and services, the exclusion of women from receiving employment offers, and the removal of protected classes from platform servers due to evaluation systems. With this backdrop in mind, this book explores the challenges the online platform economy poses to the principle of equality enshrined in both European and American antidiscrimination laws with regards to the access to goods, services, and labor markets.

The research was oriented by two main inquiries. First, it investigated the structural challenges online platforms present to the nondiscriminatory treatment of their users. Second, it examined whether antidiscrimination legal frameworks in the United States and in the European Union are equipped to address discrimination in these online spaces. After conducting case-based research, my thesis demonstrates that the structural challenges to the principle of equality rely on the aesthetic design, matching tools, evaluation systems, and the network effect of online platforms, and these aspects ultimately reinforce old biases against protected classes. Subsequently, my thesis argues that antidiscrimination laws in the United States and in member states of the European Union are only partially equipped to hold these businesses liable for discrimination occurring in their online spaces.

Regarding the structural challenges, I indicate that online platforms might enhance discrimination against protected classes. This outcome fi rst occurs by their aesthetic design choices that over-emphasize users’ protected markers before prospective transactions are concluded. Second, this outcome occurs by the design of cutting-edge matching tools that allow users to exclude protected classes from receiving goods, services, and work offers.Third, the development of facially objective evaluation systems may result in the permanent exclusion of protected classes or their poor ranking in search algorithms. Fourth, discrimination might be enhanced by the platform's network effect and its consequent impossibility and undesirability to implement prior central control over the users’ conduct.