At the fourth Cyberspace Security (Tianjin) Forum on Sept 2, 360 Group founder
Zhou Hongyi said AI has reshaped the cyber offense-defense balance and that
traditional solutions are no longer adequate. He called for an automated
offense-defense architecture centered on model-based governance—using AI to
mitigate AI risks—to secure large models, ensure reliable generated content and
trustworthy outputs. He urged full life‑cycle controls for AI agents with
auditable, intervenable and interruptible mechanisms.
