Shaping Europe's AI Selection Environment: The EU AI Act through the Frameworks of Joel Mokyr and Mancur Olson

This thesis analyses the European Union's Artificial Intelligence Act (Regulation 2024/1689) as a political struggle over the selection environment for a transformative technology. Combining Joel Mokyr's evolutionary framework of technological change with Mancur Olson's theory of collective action, it examines how three principal coalitions, incumbent technology firms, civil society and rights organisations, and workers and creators, shaped the regulation's final architecture. Mokyr's framework explains why resistance to AI regulation was intense and multidirectional, identifying the distinct categories of assets each coalition sought to protect: organisational and positional rents (Big Tech), irreversible technological trajectories (civil society), and threatened human capital (workers and creators). Olson's theory resolves why motivational intensity did not translate proportionally into political influence, demonstrating that group size, resource concentration, and selective incentives systematically advantaged concentrated industry actors over diffuse public-interest coalitions. The thesis operationalises Mokyr's aggregation rule to trace how each coalition attempted to shift regulatory decisions toward markets, designated authorities, or democratic institutions. It then assesses whether the resulting regulatory settlement exhibits features associated with Cardwell's Law, the historically recurring pattern in which innovative societies lose dynamism as distributional coalitions capture governance. The analysis contributes to the political economy of AI regulation by highlighting distributional conflict as a determinant of regulatory design.