公司治理
霸权
全球治理
现存分类群
功率(物理)
潜在Dirichlet分配
经济地理学
同性恋
政治学
全球网络
过程(计算)
计算机科学
数据科学
代表(政治)
经济体制
社会学
大数据
扩散
网络治理
全球化
透明度(行为)
国际关系
经济
复杂适应系统
区域科学
追踪
作者
Jagdeep Singh,Manpriya Dua,Amarda Shehu
摘要
Abstract Beginning with the United States in 2016, more than 70 countries and international organizations have published strategies and policy recommendations for artificial intelligence (AI) infrastructures. This article locates these policies in the shift from a hierarchical distribution of power to a flatter diffusion of power in which systemic interactions can be top-down, bottom-up or horizontal. A diffusion of power across multiple actors and regions weakens the material and socialization capabilities of hegemonic actors, resulting in global governance outcomes that are described here as ‘multiplexity’. Multiplexity offers a complex and pluralist menu of choices to actors. The computational models employed in this article show complex networks and clusters around multiplex choices that outline patterns of global governance for the evolving AI infrastructures. These networks and clusters cast doubt on many of the extant theories of global governance: those rooted in material power, wherein hegemonic states shape global governance; those where normatively motivated actors shape governance in national contexts; or those where regional patterns (North–South, East–West) are easily discernible. The article locates the origins of multiplexity in a diffusion of power entailing intersecting networks, regions, actors and world-views. There are leaders and great powers in AI, but the rest are not merely followers. In a diffused power scenario, multiple ontologies about the world coexist. The article employs big data mining, specifically latent Dirichlet allocation models from computer science, and process tracing to provide evidence of governance mechanisms for AI.
科研通智能强力驱动
Strongly Powered by AbleSci AI