Mar 12, 2026 16:00:00

Chinese state-owned enterprises and government agencies are restricting the use of OpenClaw in office environments due to growing security concerns as OpenClaw usage spreads across China.

The AI agent framework

OpenClaw , which can autonomously organize emails, write reports, prepare slides, etc. on behalf of users, has become explosively popular and rapidly adopted across China. However, this rapid adoption has led Chinese authorities to take a serious look at the risks of security and data leakage, leading the National Computer Emergency Response Technical Team (CNCERT) to issue a second warning and to begin strictly restricting its use in offices in critical sectors such as government agencies and state-owned banks.

China Moves to Limit Use of OpenClaw AI at Banks, Government Agencies - Bloomberg
https://www.bloomberg.com/news/articles/2026-03-11/china-moves-to-limit-use-of-openclaw-ai-at-banks-government-agencies

OpenClaw AI agent craze sweeps China as authorities seek to clamp down amid security fears — adoption surges as state-run enterprises are barred from use | Tom's Hardware
https://www.tomshardware.com/tech-industry/artificial-intelligence/openclaw-ai-agent-craze-sweeps-china-as-authorities-seek-to-clamp-down-amid-security-fears-adoption-surges-as-state-run-enterprises-are-barred-from-use

China issues second warning on OpenClaw risks amid adoption frenzy | South China Morning Post
https://www.scmp.com/tech/tech-trends/article/3346138/china-issues-second-warning-openclaw-risks-amid-adoption-frenzy

Unlike traditional chatbots, OpenClaw is characterized by its ability to make decisions and perform tasks on behalf of users, and it has generated such enthusiasm on Chinese social media that the phrase 'raising lobsters,' a play on its logo, has become a trending topic. Capitalizing on this boom, Tencent has introduced a compatible product called Workbuddy , and major IT companies such as JD.com are also developing their own applications. Furthermore, local governments such as Shenzhen and Wuxi are providing subsidies of several million RMB to startups that utilize OpenClaw, resulting in a heated development race involving both the public and private sectors.

An 'OpenClaw' boom is sweeping China, with AI research institutions releasing tools to support OpenClaw implementation, and long lines forming in Shenzhen for initial OpenClaw setup - GIGAZINE

However, behind this enthusiasm, Chinese authorities have expressed strong concerns about the broad data access rights and external communication capabilities that OpenClaw demands, warning that the combination of access to sensitive data, external communication, and exposure to untrusted content constitutes a 'deadly triad.'

CNCERT pointed out that OpenClaw requires high system privileges to autonomously perform tasks, creating a vulnerability where cyber attackers can steal users' system keys through 'prompt injection,' where they trick users into loading malicious commands hidden on web pages. They also reported a risk of data loss, where important emails and files could be unintentionally deleted due to user error, and warned that there have been confirmed cases where agents granted access to iMessage have gone rogue and sent hundreds of spam messages.

To address these risks, the Chinese government has prohibited employees of state-owned enterprises and government agencies from installing OpenClaw on office computers and is taking measures to quickly eliminate potential security risks. The People's Daily, the official newspaper of the Communist Party of China, has also published an interview with an expert from an agency affiliated with the Ministry of Science and Technology, warning of the dangers that AI agents pose in key sectors such as finance and energy.

According to sources, government agencies, including the country's largest bank, and major state-owned enterprises have received notices in recent days warning them to refrain from installing OpenClaw on office equipment for security reasons. Employees who have already installed OpenClaw-related applications have been instructed to report to their supervisors, undergo security checks, and remove them if necessary.

Employees of certain state-owned banks and government agencies are prohibited from installing the software not only on office computers but also on personal smartphones using the company network. This restriction is very broad, and according to one source, the same ban applies to the families of military personnel. On the other hand, not all organizations are calling for a complete ban, and some notices indicate a more lenient approach, allowing use on the condition that prior approval is obtained.

These actions by authorities highlight Beijing's growing concern that OpenClaw, which requires unusually broad access to personal data and is capable of communicating with the outside world, poses a risk of external attacks. Bloomberg reports that President Xi Jinping has long considered data a core pillar of national security, maintaining strict control through the Great Firewall censorship system while also being increasingly vigilant against data sets, including geospatial and genetic information, becoming targets for foreign actors. The report also argues that the Chinese government fears that groundbreaking technologies like OpenClaw could become deeply ingrained in society and exert influence beyond government control.

Bloomberg believes that these increased regulatory oversights will further restrict the use of unverified AI agents in government departments and state-owned enterprises in China, but also predicts that these regulations are unlikely to halt the enthusiastic boom in OpenClaw among private companies.