From Voices to Policies: Civil Society's Influence on the EU Artificial Intelligence Act (AIA)

Since its creation, the European Union has faced public backlash over concerns about democratic legitimacy. The decision-making process and initiation of legislative packages at the supranational level, involving the European Commission, the Council of the European Union, and the European Parliament, have generated concerns among scholars and European civil society, especially due to the lack of electoral accountability within the Commission. In recent decades, the Union has tried to address these concerns by evolving its consultation regime to include civil society. From business interests to NGOs and academia, numerous stakeholders contribute to EU decision-making. However, questions remain about how much these stakeholders are considered and how successful they are in attaining their preferences in specific policy fields. In the emerging field of Artificial Intelligence (AI), the Commission launched a public consultation to initiate the first global attempt to regulate AI, through innovations such as ChatGPT. Given the implications for civil society and potential international spillover, this thesis addresses how much stakeholders influenced the proposed Artificial Intelligence Act (AIA) at the policy formulation stage. As the Commission lacks substantial expertise in this new area and cannot follow other legislative designs, regulating AI offered a unique chance to evaluate lobby success between private and public interest groups. This thesis analysed the effects of (a) the nature of interests, (b) group size, and (c) information/expertise supplied by the stakeholders as potential drivers for lobby influence, inspired by collective action theory and research on lobby success. Despite the narrative of private interest domination in the Union's legislative process, the analysis of the AIA found that public interests, represented by NGOs and academic institutions, were more successful than private groups in translating their preferences into policy. Concluding that larger groups were more impactful than smaller ones, this study contradicts previous findings on the positive correlation between lobby success and information supply. Analysing the nature of interests delivers new findings on the AIA and challenges the narrative of the disadvantaged position of public interest representations in the European lobby arena.