“World-Leading?”: AI Anarchy in the UK
6
조회
0
다운로드
0
인용
0
참고문헌
초록
This paper analyses the United Kingdom’s unusual trajectory in artificial intelligence (AI) governance, positioned between the European Union’s rights-protective regulatory model and the United States’ increasingly deregulatory approach. Despite presenting itself as “world-leading” in AI, the UK has deliberately avoided creating overarching AI legislation, risk-tiered obligations, or a central AI authority. Instead, it relies on existing sectoral regulators guided by broad, non-binding principles introduced in the 2023 Pro-Innovation White Paper. This framework has been widely criticised for regulatory gaps, lack of enforcement capacity, and its failure to address discriminatory or high-impact harms arising from everyday AI systems. Following the rise of generative AI, the UK shifted political attention toward “AI safety”, launching the Bletchley Summit and establishing the world’s first AI Safety Institute. Yet its focus on hypothetical catastrophic risks has overshadowed immediate and well-documented harms, while voluntary cooperation with major US providers remains fragile. Under Labour, soft-law governance continues, although statutory regulation for frontier AI has been repeatedly delayed. In practice, the Information Commissioner’s Office (ICO) has become the UK’s de facto AI regulator, given the centrality of data protection law. The paper argues that the UK model highlights both the flexibility and the fundamental limitations of soft-law governance in addressing fast-evolving AI risks.
Abstract
This paper analyses the United Kingdom’s unusual trajectory in artificial intelligence (AI) governance, positioned between the European Union’s rights-protective regulatory model and the United States’ increasingly deregulatory approach. Despite presenting itself as “world-leading” in AI, the UK has deliberately avoided creating overarching AI legislation, risk-tiered obligations, or a central AI authority. Instead, it relies on existing sectoral regulators guided by broad, non-binding principles introduced in the 2023 Pro-Innovation White Paper. This framework has been widely criticised for regulatory gaps, lack of enforcement capacity, and its failure to address discriminatory or high-impact harms arising from everyday AI systems. Following the rise of generative AI, the UK shifted political attention toward “AI safety”, launching the Bletchley Summit and establishing the world’s first AI Safety Institute. Yet its focus on hypothetical catastrophic risks has overshadowed immediate and well-documented harms, while voluntary cooperation with major US providers remains fragile. Under Labour, soft-law governance continues, although statutory regulation for frontier AI has been repeatedly delayed. In practice, the Information Commissioner’s Office (ICO) has become the UK’s de facto AI regulator, given the centrality of data protection law. The paper argues that the UK model highlights both the flexibility and the fundamental limitations of soft-law governance in addressing fast-evolving AI risks.