OpenClaw, an AI agent tool that can execute real-world tasks like sending emails and booking flights, has become a cultural phenomenon in China despite serious cybersecurity warnings from government authorities. The tool, created in November by Austrian coder Peter Steinberger, differs fundamentally from chatbots like ChatGPT because it can take autonomous action on your behalf once granted access to your digital accounts.

What Makes OpenClaw Different From Regular AI Chatbots?

OpenClaw operates on a simple but powerful principle: users connect it to existing AI models of their choice, then give it instructions through instant messaging apps as if texting a friend. The tool can then autonomously execute tasks such as organizing files, sending emails, or booking travel without requiring manual approval for each action. This capability has captivated tech circles worldwide, but nowhere more intensely than in China, where the phenomenon has earned the nickname "lobster fever" after OpenClaw's red crustacean mascot.

Chinese entrepreneur Frank Gao, who now uses OpenClaw to manage his social media accounts, has become deeply invested in the technology. "Since January, I've spent hours on the lobster every day," Gao told AFP, referring to his OpenClaw agent. "We're family."

The adoption has been remarkably swift. Hundreds of people queued at tech giant Baidu's Beijing headquarters for an OpenClaw event where engineers helped attendees set up their "little lobsters." Similar meetups have drawn crowds from Shanghai to Shenzhen, with some municipalities, including the eastern cities of Wuxi and Hangzhou, pledging hundreds of thousands of dollars to support OpenClaw adoption and development.

Why Are Chinese Tech Companies Racing to Embrace OpenClaw?

Major Chinese technology firms have moved quickly to capitalize on the trend. Companies including Tencent, Alibaba, ByteDance, and Baidu are offering simplified installation and affordable coding plans to help users host OpenClaw agents on their cloud servers, which is considered safer than downloading the tool onto personal computers. The relatively low cost for cloud deployment in China, subsidized by these tech giants, has been a major driver of adoption.

According to Gao Rui, a senior product manager at Baidu AI Cloud, the pricing strategy is intentionally accessible: "For most people, it's likely just the price of a cup of coffee, which is why people will probably be keen to give it a try." Fear of missing out, or FOMO, is also fueling adoption, particularly among young, tech-savvy users eager to stay ahead of the curve.

The competitive landscape has intensified rapidly. In recent days, AI companies both large and small have launched their own competing agent tools, including ByteDance's ArkClaw, Tencent's WorkBuddy, and Zhipu AI's AutoClaw. This arms race reflects how seriously China's tech sector is taking the agent revolution.

What Are the Critical Security Risks?

Despite the enthusiasm, cybersecurity experts and government authorities have raised alarm bells about OpenClaw's potential dangers. The core concern centers on what happens when an AI agent gains access to your digital accounts. Once granted permission, the agent can theoretically access all services you have authorized and autonomously decide when to activate them.

"What's truly scary about agents like OpenClaw is this: once they have your digital keys, they can theoretically access all the services you've authorised, and can autonomously decide when to activate them. The attacker effectively gains a 'master key' to your digital identity," warned Frank Gao, an engineer who named his OpenClaw agent "Q" after his business name QLab.

Frank Gao, Engineer and OpenClaw User

Chinese national cybersecurity authorities and Beijing's ministry of industry and IT have issued formal warnings about OpenClaw risks. Wei Liang, an expert at the national IT research institute, advised government agencies, public institutions, companies, and individuals to "use intelligent agents such as 'lobster' with caution" in a message distributed through state media.

How Are Government Authorities Responding to the Lobster Fever?

China's approach to OpenClaw reflects a delicate balancing act between innovation and security. While municipalities are investing heavily in agent adoption and development, national authorities are simultaneously issuing cybersecurity warnings. This mixed messaging has not gone unnoticed by industry observers.

Zhang Yi, founder of tech consultancy iiMedia, characterized this approach as reflecting "the authorities' cautious tolerance towards 'lobster fever.'" The government appears unwilling to ban or severely restrict the technology, yet equally unwilling to endorse it without qualification. This cautious stance suggests that Chinese policymakers view AI agents as strategically important but potentially risky.

Steps to Safely Evaluate OpenClaw and Similar AI Agents

• Cloud Deployment Over Local Installation: Host OpenClaw agents on cloud servers provided by established tech companies rather than downloading the tool directly onto personal computers, which reduces the risk of local system compromise.
• Limit Authorization Scope: When connecting OpenClaw to your digital services, grant only the minimum permissions necessary for the specific tasks you want the agent to perform, rather than providing blanket access to all accounts.
• Monitor Agent Activity Regularly: Actively review logs and activity reports from your OpenClaw agent to detect any unauthorized or unexpected actions, treating it as you would any other account with access to your digital identity.
• Stay Informed on Security Updates: Follow official guidance from cybersecurity authorities and your cloud provider regarding best practices and emerging threats related to AI agents.

The OpenClaw phenomenon in China reveals a broader truth about AI adoption: transformative technologies often arrive faster than the security infrastructure to protect them. Peter Steinberger, the Austrian programmer who created OpenClaw to help organize his own digital life, was hired last month by ChatGPT maker OpenAI, signaling how seriously major AI companies are taking agent technology. Meanwhile, a separate team that created Moltbook, a Reddit-like platform where OpenClaw agents converse, is joining Meta.

For now, the lobster fever continues to grip China, with young users like 24-year-old college student Zheng Huimin waiting in line at tech events to experience the tool firsthand. "I'd like to give it a go to see what tasks it can actually help me accomplish," she told AFP. Yet beneath the enthusiasm lies an unresolved tension: the technology's power to automate our digital lives comes with the power to compromise them. How China and the world manage that tension in the coming months may determine whether AI agents become a trusted productivity tool or a cautionary tale about moving too fast.