Back to news
launchbytedance2026-11-04

NetEase Zhichang debuts AI-native security suite on November 4 at the Wuzhen Summit

The suite is built on the Yidun bidirectional architecture and covers content risk, deepfakes, automated black markets, model jailbreaks, sensitive data leakage and agent privilege abuse.

From November 3 to 4 the 2026 World Internet Conference Internet Light Expo was held in Wuzhen, Zhejiang, where NetEase Zhichang debuted an AI-native security product suite based on the Yidun bidirectional architecture, advancing the view that security must move from controlling what AI says to controlling what AI does. The suite received a first-release achievement certificate from the Internet Light launch hall of the expo, and NetEase Zhichang deputy general manager and Yidun business lead Zhu Haoqi gave a keynote on the AI-native security product suite built on the bidirectional AI security architecture.

Zhu Haoqi pointed out that AI has penetrated core business links including marketing, customer service, office work, research and development, knowledge bases, identity verification and transaction operations. Security governance used to revolve mainly around content, accounts and devices, but now AIGC content risk, deepfakes, automated black markets, model jailbreaks, sensitive data leakage and agent privilege abuse are the new challenges. The traditional security mindset is to defend against external intrusion, but once AI enters the business, AI itself can become a risk source.

Extending the governance object from content to behaviour is the shared direction for enterprise AI security products, with the core change being a monitoring shift from output results to execution chains and permission calls.

网易智企AI 安全易盾乌镇峰会