NetEase Zhigi Debuts AI-Native Security Product System Based on Yidun Dual-Directional Architecture in Wuzhen

Related topics top start Related topics top end Futures promotion begin Futures promotion end Miaopai begin Miaopai end Video player start Video player end Market chart begin Market chart end

Sina Tech News, evening of November 3: The 2026 World Internet Conference "Light of the Internet" Expo was held in Wuzhen, Zhejiang. NetEase Zhigi appeared at this year's conference, debuting an AI-native security product system based on the Yidun two-way architecture, and put forward a judgment: security must move from "controlling what AI says" to "controlling what AI does."

It is reported that AI for security means using AI to re-enhance security capabilities. With the help of vertical models, multi-Agent collaboration and multimodal understanding, AI can automatically analyze content and threat data, generate and verify security policies, and security personnel shift from relying on manually writing rules to acceptance confirmation of operational results.

The two directions share the same foundation, namely Yidun's long-term accumulation in the fields of content security, business risk control and security operations. On this basis, Yidun has built an AI-native security product system covering three layers, which has been recognized as a "debut achievement" and awarded a certificate by the Light of the Internet release hall.

CMA combines a self-developed content security large model with multiple types of specialized small models to build a review closed loop of "assessment—triage—disposal—feedback—optimization," supporting 7×24-hour operation and rapid horizontal expansion.

DeepGuard is based on visual understanding, moving from "looking for known forgery traces" to "understanding content and discovering unreasonable points," and integrates Deepfake detection, AI watermark recognition, silent liveness and injection attack environment detection to make multimodal joint decisions.

This capability passed the China Academy of Information and Communications Technology's "Trusted AI" face forgery detection capability special evaluation and was rated "Outstanding." The evaluation covers generative attacks, presentation attacks and injection attacks; in staged tests, the recall rate for screen re-photography and screen-pointing attacks reached 99%, with precision of 96%.

Yidun safeguards input and output through large model security fences; through Agent Guard it covers intent consistency verification, permission control, tool call risk detection and behavior interception; and through the agent security platform it achieves asset discovery, Skill/MCP supply chain security detection and sensitive data leak prevention.

Public information shows that NetEase Zhigi·Yidun is the trusted security foundation of NetEase Zhigi, founded in 2016, and has long focused on the fields of digital content security, intelligent risk control and AI security. It serves more than 400,000 developers and thousands of paying customers, and is one of the drafting units of the national standard "Basic Security Requirements for Generative Artificial Intelligence Services." news_keyword_pub,stock, iframe below the main text iframe below the main text General statement General statement QR code at the end start QR code at the end start

QR code at the end start QR code at the end start Editor name and work code

Editor name and work code