China pushes AI agents into daily services with tight controls

In short: China is moving quickly to test and roll out AI agents in public services and big consumer apps, with strong government backing and strict safety limits.

What's going on

China is positioning itself as a major testing ground for AI agents, which are AI tools that can take actions for you, like booking a ticket or filling out a form (think of them as a digital assistant that can click buttons, not just chat). A Financial Times report describes how this is being pushed by national policy and rapid adoption in cities and companies.

In March 2026, China’s annual government work report expanded the “AI Plus” plan, which encourages using AI agents in areas like manufacturing, healthcare, education, and smart cities. Several cities are also trying to attract companies building agent tools. For example, Shenzhen and Wuxi have proposed incentives and funding for local “agent hubs,” including support linked to OpenClaw, a popular framework for building agents.

Real-world trials are already happening. In Shenzhen’s Futian District, a university-built agent system is being used for public opinion analysis, cutting work that used to take weeks down to about four hours, according to reports. These pilots are often run inside isolated networks, with human review and monitoring.

What to watch

China is trying to balance speed with control. Regulators have reportedly discouraged some government and state-owned organizations from using OpenClaw due to data security concerns, favoring simpler domestic systems that are easier to audit (like using a small, clear instruction manual instead of a huge one). The key question is whether China can scale these agents across everyday services without creating privacy and security problems, especially as agents can act on a user’s behalf.

Source: Financial Times