Abstract

The widespread adoption of large AI models and generative artificial intelligence is driving a profound structural transformation of global cyberspace. Traditionally understood as an environment for information exchange among human actors, cyberspace is increasingly evolving into a "global domain of intelligent action" in which humans and AI systems jointly participate and algorithms continuously generate actions. This transformation reshapes the ontology of cyberspace along three dimensions: existence, temporality, and power. Action shifts from external initiation to endogenous generation within technological systems; time shifts from a traceable sequence of discrete events to a process of continuous evolution; and power shifts from control over access points to control over the mechanisms of action generation embedded in AI models and technological ecosystems. These changes pose systemic challenges to global cyberspace governance, which has traditionally rested on the assumptions of identifiable actors, attributable conduct, and predictable rules. They create structural challenges for the exercise of jurisdiction, rule-making, and governance enforcement. In response, generative governance calls for a shift from ex post regulation to ex ante shaping, redirecting governance from regulating conduct after it occurs toward shaping the mechanisms through which actions are generated. It also entails restructuring governance around multiple interconnected nodes and collaborative processes. Generative governance thus offers a new analytical perspective for understanding and advancing the transformation of cyberspace governance in the age of artificial intelligence.

Keywords: Artificial Intelligence; Cyberspace Governance; Generative Governance; Global Domain of Intelligent Action