China's Officials Are Using the AI Their Government Warned Them Not To

The Chinese government told its officials not to use it. They're using it anyway.

OpenClaw — a Western-developed AI agent tool — is quietly spreading across China's local governments and major tech firms, even as Beijing has issued formal warnings about its security risks. The story, reported out of Hong Kong on March 12, reads less like a tech trend piece and more like a case study in the gap between policy and practice.

The Hoops People Jump Through

Wang, a 35-year-old content director at one of China's largest tech companies, didn't stumble onto OpenClaw by accident. Getting it to run on his desktop required navigating a gauntlet: registration hurdles, payment workarounds, and setting up a dedicated "gateway" to route around China's internet restrictions. He did it all anyway.

That detail matters. People don't clear that many obstacles out of idle curiosity. They do it because the tool delivers something they can't get elsewhere — or at least, can't get as well. And Wang is far from alone.

A DeepSeek Moment, Flipped

Earlier this year, DeepSeek rattled Western markets when China's homegrown AI model demonstrated capabilities that rivaled — and in some benchmarks, beat — leading Western models. The world watched China export its AI ambitions outward.

Sponsored

Advertise with Us

[email protected]

Sponsored

OpenClaw is the mirror image. This time, a Western tool is flowing into China, spreading through the very institutions Beijing has been trying to insulate from foreign technology dependencies. The irony is hard to miss.

Chinese authorities have already flagged OpenClaw as a potential data security risk. But the warnings haven't killed the momentum. If anything, the market has responded in its own way: Chinese firms are now racing to launch OpenClaw rivals — domestically built alternatives that promise the same functionality without the geopolitical baggage.

Why This Is Bigger Than One App

The OpenClaw episode reveals something that policymakers in Beijing — and Washington — keep running into: the speed of adoption consistently outpaces the speed of regulation.

For investors and developers watching the AI agent space, the takeaway is significant. AI agent tools — software that can autonomously plan and execute multi-step tasks — are moving from novelty to utility fast enough that users in restricted environments are willing to build their own workarounds to access them. That's a signal about where the market is heading, not just where it currently sits.

For China watchers, the dynamic is familiar but worth tracking. The pattern — foreign tech gains traction, government warns, domestic clones emerge — has played out before with social media, search, and cloud services. Each time, the domestic alternative eventually dominates. The question is how long the window stays open, and what data flows through it in the meantime.

For cybersecurity professionals, the concern is concrete: when users route sensitive work data through unofficial gateways to access foreign AI tools, the attack surface expands in ways that are difficult to audit or contain.