China's Z.ai disables AI coding assistant features after security issue

BEIJING, Sept 21 (Reuters) – Chinese startup Z.ai said on Monday it had disabled some features of its flagship AI coding assistant after some users reported it was uploading entire local code repositories onto overseas cloud servers without their consent.

Beijing-based Z.ai, also known as Zhipu, ​apologised in a social media post after Chinese developers last week wrote on social media ‌that its ZCode had uploaded their code data from open-source developer platform Git to Alibaba Cloud.

Z.ai referred Reuters to its public statements when asked for further on the incident.

In a statement on Friday, Z.ai said the issue originated from ZCode’s “Codebase Indexing” feature which was enabled by default, and that it had patched the software vulnerability.

“Going forward, we will establish an ongoing ‌product ​security vulnerability reporting and response process,” Z.ai said in a separate post ⁠on its official ZCode X account ⁠on Monday.

The rare public disclosure of a security breach by a Chinese AI lab comes amid global warnings frontier AI security risks and after several leading U.S. AI companies have announced AI hacking and “rogue” AI agent incidents in recent weeks.

China’s cyber regulator released an updated AI safety framework policy ​last week, warning AI models’ shutdown resistance, evaluator deception and sandbox escape.

Z.ai said last month its GLM-5.3 model approaches Anthropic’s Mythos in finding software vulnerabilities and released it after a two-week review, becoming ⁠the first Chinese lab to explicitly delay an AI model release ⁠for safety reasons.

INDEPENDENT SECURITY ASSESSMENT

Z.ai had initially apologised on Friday and said the ​data had been deleted, but users said they found the data was encrypted with a backend private key ​held only by Z.ai, meaning they could not open or verify their own uploaded files ‌or independently confirm deletion.

Chengming Technology said on social media on Friday that six of its company coding workspaces were uploaded onto the cloud without ⁠consent by ZCode, including sensitive data such as complete source code, database passwords and employees’ personal information.

On Monday, Chengming Technology retracted its statement, saying that it had “wrong evidence”. The Chinese tech firm did not immediately respond to ⁠a request for .

Z.ai said on ‌Monday that it had open-sourced the coding assistant which runs its latest GLM-5.3 ⁠AI model and disabled certain features, pledging to make the product more ​transparent.

“We welcome ‌developers to continue reviewing ZCode and reporting potential issues.”

Z.ai said an independent ​security assessment by ⁠the Chinese industry ministry’s affiliated IT standards think tank and Chinese cybersecurity firm NSFOCUS found that users’ code data had been deleted and was not retained by the cloud platform.

Z.ai said it had enabled a zero-data retention feature on the coding assistant used by developers and tech enterprises.

“Once again, we sincerely apologize and welcome continued scrutiny from the community. The full security assessment report will be released soon,” Z.ai wrote on Monday.

(Reporting by Laurie ​Chen; Editing by Alexander Smith)