Tutor: Prof.ssa Piera Loi Co-tutor: Prof.ssa Ombretta Dessì

Keywords: anti-discrimination protection; transparency; digital transition; automated decision making; information and consultation rights; collective bargaining.

Short bio:

At the University of Cagliari in 2020 she obtained a Master’s Degree in Law, with a thesis in Comparative Constitutional Law on the underrepresentation of women in elected assemblies. Later, in 2021, she completed a Level II Master’s Program in Industrial Relations in the Public and Private Work, presenting a thesis on pay transparency as a tool to combat the gender pay gap. Since 2023, she has been a PhD candidate in Labour Law at the University of Cagliari, with a research project focused on algorithmic discrimination arising from the use of artificial intelligence in the workplace.

Thesis abstract:

The aim of this research project is to analyze measures for preventing and combating discrimination, both gender-based and those based on other protected factors, within the employment relationship. It specifically focuses on the effectiveness of current protection mechanisms in addressing forms of discrimination resulting from the growing use of algorithmic systems and artificial intelligence in supporting or replacing managerial decisions. The question regarding the discriminatory potential of algorithms has long been answered affirmatively: the negative consequences arising from the use of algorithmic decision-making systems disproportionately affect individuals who carry protected factors within society. Through the study of such systems and the existing protection solutions in the legal framework, both national and supranational, the research aims to establish how discrimination within the workplace may result from unjust algorithmic management. In order to understand the challenges facing anti discrimination law and its effectiveness, the first goal is to define the concept of algorithmic discrimination, exploring how algorithmic models can enable discriminatory biases, both directly and indirectly. The source of the discriminatory risk, in fact, lies not only in the human factor, which is crucial for the development and operation of algorithmic software, but also, and above all, in the opacity and consequent unpredictability of non-deterministic algorithms. One of the main objectives of the research is to critically examine the regulatory responses, in order to evaluate, on the one hand, the effectiveness of the traditional model that underpins the logic of the legislation on the matter—foreseeing mandatory rules and prohibitions, accompanied by post facto protection measures—and, on the other hand, to analyze, through the necessary integrated multidisciplinary approach, the new perspective adopted at the European level, which favors a system of identifying and assessing discriminatory risk ex ante, accompanied by preventive measures and risk minimization. The research also intends to explore the need to further expand the discussed guarantees by promoting a greater contribution from collective bargaining. Although often perceived as a slow process, collective bargaining proves to be a crucial tool in bridging the gap between the technological innovation that characterizes productive contexts and the legal protection of workers against discrimination. Finally, the project aims to explore the role of the Labor Inspectorate in monitoring companies, suggesting the integration of new inspection and investigation methodologies that are more effective in combating algorithmic discrimination.