Tencent has begun a nationwide campaign to install its new AI assistant, nicknamed “Longxia” (literally “lobster”), on consumer devices and local services free of charge. The company is pushing the agent into homes and small businesses with on-site install services and online booking that local outlets report are fully booked for days, a sign of unexpectedly brisk demand.
Longxia is being promoted as a hands-on automation tool: it can directly operate computers, draft code and content, and — the company claims — assemble a working website in roughly twenty minutes. That practical pitch has fed a viral “install-and-try” culture across dozens of Chinese cities, where local service desks even require users to sign safety responsibility agreements before technicians proceed.
The rollout, however, has collided with a rapidly mounting set of concerns. China’s Cyberspace Administration has publicly named Longxia as posing “serious security risks,” while media reports and watchdogs have documented early privacy leaks and at least some users reporting negative consequences after installations. Security entrepreneurs and industry veterans are warning that the agent’s broad device permissions and autonomy create attack surfaces that have not yet been systematically hardened.
The product has also become mired in intellectual-property friction. The founder of an independent project called OpenClaw accused Tencent of copying his work; Tencent replied that it hopes the ecosystem will continue to thrive and sought to downplay the dispute while pressing ahead with installs. Commentators inside the industry — including prominent figures who have urged tighter controls over platform permissions — say the episode highlights the speed at which large firms can push novel, powerful software into millions of endpoints.
Local governments and some business groups have actively encouraged adoption, framing Longxia as an economic productivity booster that can democratize website building and simple automation. That encouragement has helped the program spread quickly, with scenes of long queues and “Longxia” demo events reported in multiple municipalities. Still, the state’s recent warnings and individual reports of privacy exposure have prompted a rethink among consumer-rights groups and some municipal IT departments.
For Tencent, the free-install drive serves multiple purposes: accelerate user adoption, collect usage data, and establish Longxia as a de facto standard for lightweight automation on consumer devices. But those same aims raise questions about consent, data governance and the concentration of control. Industry voices are calling for clearer rules for what automated agents are allowed to do on end-user machines and for standardized security audits before mass deployment.
The unfolding episode is a microcosm of a larger struggle inside China’s technology sector: companies race to entrench AI-driven services in everyday life while regulators and civil society scramble to define acceptable bounds. How Beijing balances the economic benefits of rapid, widespread deployment against the political and technical risks will determine whether Longxia becomes a mainstream productivity tool or a case study in premature scaling.
Ultimately, the Longxia rollout is not just a product launch but a stress test of governance, market power and societal tolerance for automated systems that operate close to users’ data and devices. The coming weeks will likely bring more scrutiny, possible technical restrictions on permissions, and pressure on firms to disclose how these agents make decisions and handle personal information.