Towards Adaptive Categories: Dimensional Governance for Agentic AI

The article presents a framework for "dimensional governance" in AI, addressing the inadequacies of traditional categorical governance as AI systems transition to dynamic agents. It emphasizes the need to dynamically track decision authority, process autonomy, and accountability (the 3As), allowing for pre-emptive adjustments to risks as systems evolve. This approach promotes adaptable categorization that can evolve with AI capabilities, providing a more resilient governance model that responds to the complexities of human-AI interactions.