China has warned state agencies and major banks against installing the open source AI agent OpenClaw, citing potential data exposure and cybersecurity risks linked to autonomous tools with deep system access.
Chinese authorities have restricted state-run enterprises and government agencies from installing the open-source autonomous AI agent OpenClaw on office computers, highlighting rising regulatory scrutiny of open-source AI tools with deep system access.
Notices were reportedly sent to government departments and some of the country’s largest banks warning against installing OpenClaw applications. Agencies were instructed to report existing installations, conduct security checks, and remove the software where necessary.
The restrictions extend beyond official devices. Employees at state-run banks and government bodies have reportedly been barred from installing OpenClaw on office computers as well as personal phones connected to company networks. The warning has also been extended to relatives of Chinese military personnel. In some cases, the notices do not impose a complete ban but require prior approval before using the tool.
Authorities are concerned about the security implications of autonomous AI agents that require broad access to private data and can communicate externally. Cybersecurity experts have warned that such systems may be vulnerable to prompt injection attacks, malicious plugins, data leaks, supply-chain attacks, and other risks stemming from deep system access.
Despite these concerns, OpenClaw has rapidly gained traction among developers and consumers in China. Installation drives in Shanghai have attracted strong participation, including users aged 60 and above experimenting with agentic AI tools.
Major technology firms including Tencent and JD.com are building services around the platform, while Alibaba, Tencent and Baidu are supporting developers with cloud infrastructure.
Following reports of the restrictions, shares of companies developing OpenClaw-based tools reacted sharply. Tencent surrendered earlier gains, while MiniMax shares declined by more than one per cent.
The episode underscores growing tension between open-source AI innovation and national security concerns, even as China’s startup ecosystem experiments with new applications ranging from AI-agent social networks to automated recruitment and influencer management platforms.
