Governing Public-Sector AI: A Comparative Study of Saudi Arabia and the United Kingdom

Abstract

Artificial intelligence (AI) has moved from a largely technical domain to the centre of global legal and policy debates . Today, AI systems include search engines, recommendation systems, and an emerging class of generative and “frontier” models that can produce text, code, images, and other content at scale. These systems create opportunities for economic growth, efficiency, and public- sector innovation, but they also give rise to complex risks for safety, fundamental rights, and security. This dissertation focuses specifically on the use of AI by government and public-sector bodies rather than attempting to cover all private-sector applications. States and international organisations have begun to develop principles and legal frameworks to govern these risks under the banner of “trustworthy” or “responsible” AI. Saudi Arabia and the United Kingdom occupy distinctive positions within this emerging global landscape. Saudi Arabia has framed AI as a core enabler of its Vision 2030 economic diversification agenda and has created a centralized institutional structure through the Saudi Data and Artificial Intelligence Authority (SDAIA), supported by the National Data Management Office (NDMO) and other digital regulators. The National Strategy for Data and AI (NSDAI) sets ambitious targets for AI-driven growth and capacity building. Moreover, SDAIA has adopted national AI ethics principles and guidance intended to shape both public and private sector deployment . By contrast, the UK has presented itself as a global leader in “pro-innovation” AI governance. The 2021 National AI Strategy and the 2023 White Paper “A pro-innovation approach to AI regulation” articulate a model that avoids a single comprehensive AI statute and instead relies on existing regulators applying a common set of AI principles within their respective mandates. The establishment of a dedicated AI safety body, the AI Security Institute, signals an increasing focus on advanced AI safety and international cooperation, even in the absence of a full AI Act . Both jurisdictions seek to be AI leaders through different institutional designs and regulatory techniques. This dissertation uses these differences to examine how states can pursue innovation while managing the risks and societal impacts of AI.