CHINESE firms have recently unveiled two advanced AI models aimed at enhancing vulnerability discovery, raising alarms among cybersecurity professionals. Notably, Zhipu AI's GLM 5.2 model has shown superior performance in vulnerability detection benchmarks compared to leading models from US companies. Another model by 360 Security Technology has reportedly identified thousands of vulnerabilities already.
Experts stress that defenders need to promptly address security weaknesses within their infrastructures to keep pace with increasingly sophisticated AI-driven attacks, which have drastically reduced the time required to exploit vulnerabilities. Furthermore, the availability of open-weight models allows both attackers and defenders to operate more flexibly, underscoring the importance of adopting effective integration of AI technology in cybersecurity practices.