Taking the lead in taking action against AI-managed accounts, why does Xiaohongshu choose to fire the first shot?

The real experiences and emotions that underpin the atmosphere of the Xiaohongshu community can only come from real people. 

While OpenClaw has gone viral and the security of "raising lobsters" remains questionable, AI-managed accounts created with the help of AI tools like OpenClaw have quietly spread to mainstream content communities.

Against this backdrop, Xiaohongshu has taken the first action against AI-managed accounts, marking the first time a mainstream platform has explicitly regulated the abuse of AI agent tools in content communities.

As a UGC community founded on real-person sharing, sincerity and authenticity have always been the core foundation of Xiaohongshu. When AI Agents attempt to erode the authenticity of the community, they have crossed the uncrossable bottom line of the platform.

This move by Xiaohongshu is not only a protection of its own ecosystem, but also a timely correction of industry chaos.

 The Action Taken

OpenClaw, an open-source AI agent, has sparked a global boom recently. As its icon looks like a crayfish, the term "raising lobsters" has quickly become a hot topic in the tech circle.

Different from traditional AI assistants that can only carry out conversations, the core advantage of the "lobster" is that it "can take practical actions" — just like a personal super assistant, OpenClaw can not only interact with users and complete tasks according to instructions, but also "take over" the user's computer, automatically call different data to complete complex tasks, and stay online 24/7.

To a certain extent, the "lobster" has realized the technological leap of AI from conversation to execution. In other words, it lowers the threshold for people to understand how to use Agents, and people can directly manipulate computers to complete relevant operations only with natural language instructions. Soon, the "lobster craze" quickly spread out of the tech circle, and topics such as how to install and raise the lobster once became the most discussed topics on social media.

But just a few days later, voices about the potential risks of information leakage and other hazards brought by the "lobster" began to emerge. On March 10, the National Internet Emergency Center of China released a risk warning on the safe application of OpenClaw, noting that the default security configuration of OpenClaw is extremely fragile, and once attackers find a breakthrough, they can easily gain full control of the system.

This means that for ordinary users without professional capabilities, once the vulnerability of OpenClaw is maliciously exploited by cyber attackers, it is no different from giving the key to their own house to strangers on the street.

While the security of the "lobster" is highly controversial, AI-managed accounts formed with the help of AI tools such as OpenClaw have quietly penetrated into all major mainstream content communities, and Xiaohongshu is one of them.

What is an AI-managed account? In short, it refers to accounts operated under the AI hosting mode, which automatically generate content, publish notes, and simulate real-person interactions in scenarios such as comments, private messages, and group chats through technical means. For example, the Xiaohongshu account "Cyber Crab" has completed all work including persona setting, content creation and daily operation with AI. Although it simulates real-person expression, the content has a high degree of homogeneity and its authenticity cannot be verified, and it has been banned at present.

At present, Xiaohongshu has officially taken action against these batch-operated AI-managed accounts.

On March 10, 2026, Xiaohongshu released a governance announcement on cracking down on AI-managed operation accounts, strictly prohibiting any behavior that uses technical means to simulate real people, generate non-real content or carry out false interactions, and opened a reporting channel, where users can report suspected AI-managed accounts through the reporting function on the note or account page.

The announcement also mentioned the specific governance measures implemented for AI-managed accounts:

1. For ordinary accounts that occasionally use AI hosting to write and post notes or conduct false interactions, the platform will take gradient handling measures such as warnings and restricting content distribution according to the severity of violations.

2. For accounts that complete registration, publishing and interaction directly through AI hosting tools, or accounts where all public notes on the homepage are posted by AI hosting, the platform will ban them.

As for how to judge whether an account is an AI-managed account, a source close to Xiaohongshu told 36Kr that the platform will comprehensively determine whether the account is automatically manipulated and published by AI according to the publishing behavior of the account, and also retain the appeal entrance.

This move by Xiaohongshu is not only the first shot fired by a mainstream content platform against AI-managed accounts, but also marks the first time that a mainstream platform has explicitly regulated the abuse of AI agent tools in content communities.

 Authenticity as the Bottom Line

Why did Xiaohongshu choose to fire the first shot?

The reason is not complicated — the existence of AI-managed accounts is contrary to the authenticity that Xiaohongshu has always advocated. As a UGC community founded on real-person sharing, "sincerity and authenticity" are the foundation of Xiaohongshu, and UGC content derived from real people's experiences is also the most core asset of Xiaohongshu.

The community atmosphere with a vivid sense of real people once allowed Xiaohongshu to carve out a niche in the fierce internet competition and achieve what it is today. The "human touch" behind it all comes from users' real sharing and expression. Xiaohongshu officials once disclosed that from January to November 2025, more than 70 million comments were generated on Xiaohongshu every day, and about 200 million users sought purchase advice on Xiaohongshu every month.

However, the logic of AI-managed accounts is to pursue scale realization. Once the content template is verified to be effective, it will be copied in batches, and finally realize monetization by receiving advertisements or diverting traffic to product links. It has nothing to do with real-person creation, with worrying authenticity, let alone producing high-quality content.

Once a large number of such accounts settle in, they will inevitably gradually erode the core assets of Xiaohongshu. The low-quality and homogeneous content generated by AI will occupy the distribution space of real UGC content and dilute the weight of high-quality content. More importantly, when the subtle "AI flavor" slowly seeps into the community, over time, users will find it difficult to distinguish whether there is a real person or an algorithm behind the screen, which will eventually affect users' trust in the platform and shake the foundation of the community.

Just as the community is the soil for all the value of Xiaohongshu, UGC content is still the foundation of Xiaohongshu's commercialization, which is also the reason why the "grass-planting" (product recommendation) culture was born and verified on Xiaohongshu. Today, users trust the content on Xiaohongshu more and are more likely to generate consumption behaviors, which essentially lies in the fact that users are more willing to trust product recommendations from ordinary people, rather than pervasive advertisements.

With the development of the grass-planting culture, Xiaohongshu has become the last stop for people to make consumption decisions. But AI-managed accounts have spawned a series of gray industrial chains. Once these accounts disguise themselves as ordinary users' notes to divert traffic and realize monetization, the content that is hard to distinguish between true and false will not only damage users' trust, but also seriously disrupt the normal operation of Xiaohongshu's brands and merchants.

Since the first half of last year, in order to govern false marketing content on the platform, Xiaohongshu has specially set up an "anti-false marketing team" internally, and banned more than 12 million false accounts. Those accounts were operated by real-person teams at that time. If AI-managed accounts are used, the number will increase exponentially. Cracking down on AI-managed accounts is also a kind of "intervention in advance" for Xiaohongshu. 

Example of AI-managed account

In the final analysis, this governance action of Xiaohongshu is a joint protection for users, creators and the platform — for users, Xiaohongshu guards the space for real sharing and prevents users from being misled by false content; for creators, it maintains a fair and healthy creation environment, so that high-quality real-person content can form a positive cycle and get the due exposure and return; for the platform, it adheres to the core value of the community and retains its own foundation for survival.

 Where is the Boundary of AI Agent?

It is worth noting that Xiaohongshu's crackdown on AI-managed accounts this time does not mean a blanket ban on the use of AI technology.

In fact, AI Agent is not a monster. In community scenarios, its reasonable application can provide assistance for users and creators, and enrich the content ecology of the community.

What Xiaohongshu bans this time is the behavior that completely relies on AI hosting to publish, interact and even operate accounts, rather than restricting creators to use AI technology as an auxiliary tool in creation. In other words, what the platform really values is who is behind the content production, a real person or AI.

After all, those real experiences and emotions that underpin the atmosphere of the Xiaohongshu community can only come from real people. The emotional flow between people cannot be replaced by any AI.

Therefore, when AI Agent attempts to erode the authenticity of the community, it has crossed the uncrossable bottom line of Xiaohongshu — technology can serve content creation, but cannot be used to forge authenticity and deceive users.

The essence of a community is "real connection between people". The value of technology is to assist this connection, rather than replace it. People hope that AI is a tool to assist life, but they do not want it to seep into their lives without a trace, at least not to be the role that decides what they see, induces, deceives, and even affects their decisions. AI Agent can be used as an assistant, but it can never replace real people to become the main body of the community.

Especially for UGC communities like Xiaohongshu that rely on trust relationships, what users pursue is sharing of real experiences, sincere interaction and the warmth between people, rather than false content carefully packaged by algorithms, let alone "template notes" produced in batches by AI. Once this authenticity is broken, the cohesion of the community will be dissolved, and the value of the platform will collapse accordingly.

From a macro perspective, the large-scale implementation of AI Agent is an inevitable trend. To a certain extent, AI and communities are not in an antagonistic relationship. Only by finding a reasonable application boundary can symbiosis be realized. Xiaohongshu's governance of AI-managed accounts is not only a protection of its own community ecology, but also provides inspiration for the industry on the application of AI Agent technology in community scenarios.

After all, dynamic symbiosis with AI will be a future test that all content platforms have to face. Today is just the beginning.