In the era of generative AI, the underlying logic and application scenarios of network information content have undergone new transformations, spawning an entirely new human-machine hybrid information environment. Network information content risks are confronted with the dual challenges of "disorderly growth" and "unbounded deviation" of AI-Generated Content (AIGC). This paper constructs a novel risk framework for network information content under the influence of generative AI technology, integrating both informational and systemic dimensions, and further develops and analyzes an AIGC risk pyramid model. Addressing these evolving risk patterns, the paper proposes a risk governance pathway for network information content in the era of generative AI based on agile governance principles, covering three key levels—governance philosophy, governance subjects, and tools—providing references for the future development of network information content ecosystems.