Global Markets
Home›Global Markets›China›Zhipu AI unveils GLM-5.3 and open-source cybersecurity…
Zhipu AI unveils GLM-5.3 and open-source cybersecurity tools
The company says it will use a layered risk-review system for GLM-5.3 and reserve its most sensitive offensive capabilities for verified users.
Zhipu AI, now known internationally as Z.ai, announced its GLM-5.3 model and a new “Shield of Open Source” initiative aimed at cybersecurity, according to SCMP Economy.
The program includes free security audits to help users patch software vulnerabilities, along with automated code-auditing tools through ZCode, the outlet reported.
The initiative comes as the global debate intensifies over whether open-weight AI models can increase cyber risks by making it easier for attackers to access advanced capabilities.
Zhipu said it will apply stricter oversight with a layered risk-review system that blocks high-risk requests, while reserving the model’s most sensitive offensive capabilities for verified users under a restricted access plan called “Cybersecurity Trusted Access,” SCMP Economy added.