By Da Cheung
On March 10, lifestyle-focused social media and e-commerce app Rednote announced a crackdown on accounts operated entirely by artificial intelligence. The platform, known in Chinese as Xiaohongshu, said it will permanently ban accounts that use AI tools to register, automatically generate posts, publish notes, and simulate human interactions in comments, private messages, and group chats. Ordinary users who occasionally use AI to draft or post content will be given warnings and traffic to their accounts will be restricted.
Rednote, owned by privately held Xingyin Information Technology, has gained massive international traction since it started its overseas expansion in early 2025 and is often described as ‘Instagram meets TikTok.” It was the most-downloaded app on Apple’s App Store in the U.S. in January 2025, fueled by a surge of users seeking an alternative to TikTok, although for the year as a whole , it was only the 20th most-downloaded.
As of August 2025, the platform claimed to have over 350 million monthly active users, according to a Chinese media report citing the company, most of them young women.
Rednote’s administrators justified the platform’s decision by saying that because it relies heavily on users sharing authentic travel, shopping, and lifestyle experiences, AI-generated content threatens the “warm, underlying tone of the community.”
The rise of the digital user
The sudden bans are a direct response to a viral trend in China involving an open-source AI agent framework called OpenClaw, created by Peter Steinberger. Nicknamed “Lobster” by Chinese netizens, OpenClaw differentiates itself by not just offering advice, but by actively executing multi-step tasks across platforms.
Instead of dealing with complex programming, users simply give OpenClaw natural language instructions, effectively hiring it as a 24-hour digital employee. With a basic conversational prompt, the AI can independently brainstorm topics, draft engaging copy, generate images, schedule posts, and reply to follower comments. To circumvent platform detection, developers have integrated browser tools that mimic natural human behavior — such as pausing organically while scrolling.
For many content creators, this technology transforms social media marketing from a creative grind into an automated assembly line. One user reported handing a dormant lifestyle account over to the AI with just a single sentence of instruction. The AI independently researched trends, scheduled content, and boosted account engagement by over 300% — requiring only 15 minutes of human oversight for the entire week.
The “raising lobsters” phenomenon has caught the attention of the state, the private sectors, and society as whole. Major internet conglomerates, including ByteDance, Tencent, and Alibaba, are rushing to follow up on similar automated agent technologies. Meanwhile, various local governments have rolled out policies to encourage AI development, offering huge subsidies to startups leveraging OpenClaw.
A global social media dilemma
While Rednote is actively suppressing fully automated accounts to avoid platform manipulation, Meta is exploring the boundaries of AI integration in contrasting ways.
Last December, Meta was granted a patent outlining how a large language model can simulate a person’s social media activity. “The language model may be used for simulating the user when the user is absent from the social networking system, for example, when the user takes a long break or if the user is deceased,” according to the patent authored by Meta’s Chief Technology Officer Andrew Bosworth, as reported by Business Insider.
While a Meta spokesperson told Business Insider that the company has “no plans to move forward with this example,” the contrast between Rednote’s crackdown and Meta’s patent filings underscores a broader challenge for the industry: how the global social internet will balance the hyper-efficiency of AI automation with the authentic human interaction users want and expect from social platforms.
