Non-State Actor Preferences in AI Governance - Analysing the 2021 Public Consultation by the Council of Europe on a Legal Framework for Artificial Intelligence

Sammanfattning

This thesis aims to contribute to understanding the preferences of non-state actors in the global governance of AI. It distinguishes between business, academic, and civil society actors and analyses their overall regulatory preferences as well as their preferences in four specific areas of AI regulation. It derives theoretical expectations from existing literature based on the rationalist theory of preference formation and tests these expectations by applying OLS regression, using data from a public consultation procedure conducted by the Council of Europe in 2021. The results show that all actor types want a relatively high level of regulatory stringency for AI. Still, there are significant and substantial differences between actor types. Civil society actors want the most stringent regulation, and business actors want the least stringent regulation. Contrary to expectations, the preferences of academic actors align with the preferences of business actors rather than the preferences of civil society actors. Both business and academic actors want significantly less stringent regulation than civil society actors for all regulatory areas. The preferences between the regulatory areas only differ slightly. Civil society actors have more homogenous preferences than business and academic actors. This thesis adds nuance to the existing research. It supports but also challenges existing assumptions and highlights the complexity of stakeholder dynamics in AI governance. Understanding stakeholder preferences is essential for developing effective and inclusive regulatory frameworks that balance innovation, societal benefits, and ethical considerations in the rapidly evolving landscape of AI technologies