China fuels OpenClaw adoption despite mounting security concerns

Local governments in China are offering subsidies to drive adoption of the OpenClaw AI agent, even as security experts in China and abroad warn that improper configuration of the software could carry serious cybersecurity risks.

Chinese technology hubs are rapidly moving to promote the open-source AI agent framework through financial incentives, startup programs, and policy initiatives aimed at accelerating the development of applications built on top of the platform.

Districts in cities including Shenzhen, Wuxi, Longgang, and Hefei have unveiled draft measures to foster local ecosystems around the technology. This is part of the country’s broader national plans to expand artificial intelligence across industries under its “AI plus” strategy.

The Wuxi high-tech district is offering up to 5 million yuan (about $690,000) ​for projects applying OpenClaw to manufacturing-related technologies, while both Longgang and Hefei have proposed subsidies and financing of up to 10 million yuan (about $1.4 million) for companies that build notable OpenClaw applications.

OpenClaw, created by Austrian developer Peter Steinberger, has quickly become one of the fastest-growing projects on GitHub since its release late last year. Unlike conventional chatbots, the AI agent can execute actions directly on the user’s behalf, such as sending emails, organizing files, scheduling meetings, and more.

The technology’s ability to automate digital work has sparked intense interest in China. In one example highlighted by regional media, nearly a thousand people queued outside Tencent’s headquarters in Shenzhen to have engineers install OpenClaw on their computers for free.

Tread with caution

Despite the surge in enthusiasm, regulators and cybersecurity experts have warned that the same capabilities that make autonomous AI agents powerful also introduce new risks.

China’s National Vulnerability Database, a cybersecurity information center under the Ministry of Industry and Information Technology, has cautioned that improperly configured OpenClaw deployments could expose users to cyberattacks or privacy leaks, because the software often requires broad system permissions to operate effectively.

Security researchers have also raised concerns about the broader attack surface created by AI agents. In fact, Microsoft has explicitly warned that users shouldn’t run OpenClaw on normal PCs.

Researchers have also reported thousands of publicly exposed OpenClaw instances online, some vulnerable to takeover through remote code execution flaws. Other incidents have shown how autonomous agents can behave unpredictably when given extensive system privileges.

The fact isn’t entirely lost on Chinese policymakers. Draft policy measures in Wuxi, for instance, call on cloud platforms offering OpenClaw to block access to sensitive data directories and to explore creating AI compliance service centers focused on cross-border data transfers and intellectual property protection.

Unlock more exclusive Cybernews content on YouTube