AI Governance and Digital Rights Protection: Insights from the OECD Dataset

• In Event: The Global Governance of Artificial Intelligence

Abstract

What are the present trends in AI governance, and how do they reflect different approaches to the protection of digital rights? The misuse of AI, such as invasive data collection and discriminatory algorithms, has posed significant threats to privacy, non-discrimination, and other fundamental freedoms. Utilizing the OECD dataset of national AI policies, this paper coded variables covering 885 policies from 70 countries and the EU. The focus on digital rights protection encompasses two dimensions: target actors, including the private sector and government agencies, and the costs of regulation, which range from informal to formal governance. The newly compiled dataset includes variables such as technical standards, types of legal tools, the strength of regulatory institutions, and the extent of policy implementation and enforcement. Combining a keyword-based approach for analyzing policy documents and hand-coded digital rights indicators, this paper employs topic modeling, network analysis, scaling analysis, and regression analysis. The findings suggest that the EU and several AI superpowers have been influential in setting the agenda for AI ethical norms and governance models. However, distinct regional patterns have emerged regarding the regulation of firms and government agencies, with reactive countries strategically adopting AI policies to serve their domestic interests. By creating a systematic taxonomy of AI regulation, this paper offers a clear theoretical conceptualization and detailed observational analysis to examine the variations in AI governance across countries and over time. This paper speaks to the debate on the changing role of the state in the digital age and the balancing of public goods with special interests.

Author

• Huimin Li, University of Texas at Austin