Xiaohongshu (小红书) Refuses AI

Ban targets AI "proxy operation" as a threat to authenticity

Xiaohongshu (小红书) has announced a crackdown on accounts run by AI agents, saying it will directly ban accounts that use automated systems to generate posts, reply in comments and private messages, or otherwise simulate real users. The move follows the rapid spread of an AI agent nicknamed "OpenClaw" (commonly called “龙虾”), which, it has been reported, can take a single natural‑language instruction and draft titles, generate images and text, schedule posts and even leave comments — effectively running an account end‑to‑end. Reportedly this is the first time a major Chinese content platform has explicitly red‑carded the “agent as proxy operator” model.

Why Xiaohongshu is different — and why this matters

For Western readers: Xiaohongshu began as an overseas‑shopping guide and has evolved into a lifestyle and recommendations community where perceived authenticity — the “living person” voice — underpins commercial value. The platform says it now has more than 350 million monthly active users, with roughly 9 million notes posted daily and very high engagement metrics. Brands have leaned on that trust to convert recommendations into purchases; that is the platform’s core competitive advantage. So when machine‑generated content starts to flatten voice and scale cheaply, what happens to a business built on real human experience?

Homogenization, creator harm and enforcement data

Platform operators argue AI output tends to be polished but generic: the voice flattens, the small imperfections and emotional texture that signal a human author disappear, and comment‑section warmth can become mechanical. Xiaohongshu’s own governance figures, disclosed by the company, say it removed some 600,000 low‑quality AIGC notes in the first half of 2025 and banned more than 10 million accounts tied to batch fabrication. The risk is systemic: if algorithmic feeds reward cheap, mass‑produced posts, genuine creators will lose reach and incentives, and the “seed‑planting” (种草) economy built on trust could be hollowed out.

A business balancing act in a global AI environment

The policy is also a commercial calculation. Xiaohongshu has been experimenting with multiple monetization paths — external links, a marketplace, local services and paid notes — but the platform’s commercial value depends on authentic engagement. The crackdown signals a choice to protect that asset even as AI tools proliferate. And it comes against a broader backdrop of intense global debate over AI governance: platforms in China, like their Western peers, face pressure to police disinformation and fraud while navigating national AI strategy and international technology tensions. Will policing proxy agents preserve Xiaohongshu’s soul, or merely slow an inevitable shift in how online influence is produced? The answer will shape both the platform’s future and the evolving rules for AI in social media.