Chinese AI tools Trae and Qoder gain after Anthropic back-door alert

This week, an alert was released by China’s National Vulnerability Database (NVDB), part of the Ministry of Industry and Information Technology, indicating that several iterations of Anthropic’s Claude Code tool have a security back door. This vulnerability could potentially transmit user identities and locations to external servers without permission. The NVDB has recommended that local organizations either uninstall the compromised versions or upgrade to the updated releases. In reaction, domestic technology companies are increasingly turning to local alternatives like ByteDance’s Trae and Alibaba’s Qoder. Cai Peng, a cybersecurity partner at Zhong Lun Law Firm in Beijing, anticipates that more Chinese firms will discontinue the use of foreign AI tools due to rising security issues and the nation's goal for technological self-sufficiency.