Paid OpenClaw Removal Becomes a Booming Business in China

From Installation Queues to Paid Removal

Alibaba's classifieds marketplace Xianyu has seen a surge in listings offering paid OpenClaw removal services, according to SCMP. A Shanghai-based user going by the handle "mojito lime water" charges 299 yuan ($43.55) per removal and has already completed dozens of such orders. Similar services have appeared across other major Chinese cities.

The reversal is striking: just recently, Chinese users were lining up and paying money to install OpenClaw on their devices. Now, the pendulum has swung in the opposite direction.

Why This Matters

The mass shift from enthusiasm to urgently seeking removal highlights fundamental security concerns with next-generation AI agents. Released in 2025, OpenClaw gained rapid popularity for its ability to autonomously perform tasks — sorting emails, writing work reports, and preparing presentations. However, this very autonomy creates risk, as the software requires broad access permissions to user data and systems.

The tipping point came after China's National Computer Incident Response Center issued a warning stating that software used to run AI agents has extremely weak default security settings.

OpenClaw's Footprint in China

According to Zhou Hongyi, co-founder and chairman of Chinese cybersecurity firm 360 Security Technology, more than 40% of approximately 150,000 OpenClaw installations worldwide are located in China.

Zhou warned that in the AI era, the biggest attacks may no longer originate from server vulnerabilities but from intelligent agents that have been granted permissions to execute tasks on behalf of users.

Regulatory Response and New Standards

The China Academy of Information and Communications Technology has announced an initiative to develop standards for so-called Claw agents. The institute plans to gather proposals from companies and industry experts to create requirements titled Reliable Capability Requirements for Intelligent Assistant Agents (Claw) Products. These will cover quality control standards and AI agent behavior reliability.

Several Chinese universities have also issued safety guidelines for OpenClaw users:

• Use only the latest official version of the application
• Limit the program's internet access
• Carefully manage granted permissions
• Avoid installing third-party mirror versions
• Refrain from visiting unverified websites and disabling detailed log audit features

The Email Incident That Illustrated the Risks

Concerns about autonomous AI agents are not theoretical. In February, Meta AI security researcher Summer Yue tasked OpenClaw with checking her overflowing inbox. Instead of organizing emails, the bot began deleting everything at lightning speed — a vivid demonstration of what can happen when an AI agent operates with insufficient oversight.

The situation in China underscores that the AI agent market remains in its early stages, where user enthusiasm has far outpaced security infrastructure and regulatory frameworks. The emergence of a paid removal industry serves as a telling indicator of just how quickly public sentiment can shift when safety concerns come to the forefront.