Xiaohongshu (RedNote) has implemented strict restrictions on AI-generated content, putting international fashion brands that rely on AI for content production in a difficult position. The platform treats authentic user experience as its core value proposition, and AI-generated images, videos, and text undermine the credibility that makes XHS valuable for Chinese consumers. Understanding exactly what is banned, what is still allowed, and how detection works is now essential for any brand managing an XHS presence.
What XHS Bans: AI-Generated Images and Deceptive Content
XHS specifically targets AI-generated images used without disclosure. This includes product photos generated by tools like Midjourney, Stable Diffusion, or DALL-E that are presented as real product photography. It also includes AI-generated “lifestyle” images that simulate genuine user experiences. The platform has deployed AIGC (AI-Generated Content) detection systems that flag images exhibiting characteristic AI generation patterns. In 2025, XHS updated its community guidelines to require explicit disclosure when any AI-generated visual element appears in a post, with non-disclosed AI content subject to content removal and account penalties.
Why XHS is Taking a Harder Line Than Other Platforms
The reasoning is commercial. XHS positions itself as a trusted discovery platform where real users share genuine product experiences. Its advertising model depends on consumers believing that the content they see reflects authentic opinions. If AI-generated content becomes indistinguishable from real user posts, the platform loses the trust differential that makes it more effective for product discovery than traditional advertising. Unlike Douyin, which is entertainment-first and more tolerant of produced content, XHS is built on the assumption that what users see reflects what real people actually bought and used.
How XHS Detects AI-Generated Content
XHS uses multiple detection methods. Image analysis tools identify visual patterns characteristic of diffusion model outputs: unnatural skin textures, inconsistent lighting, deformed hands and fingers, and artifact patterns in backgrounds. Text analysis flags posts written in patterns typical of large language model outputs: overly structured format, absence of personal voice, and language that matches AI writing patterns in Chinese. User reporting also plays a role: XHS users actively flag content they suspect is AI-generated, and the platform acts on high-volume reports. Accounts with repeated AI content violations see progressive penalties from reduced reach to account suspension.
What Fashion Brands Can and Cannot Do on XHS
AI tools used for internal planning, caption drafting, or translation are not the issue. The restriction is specifically on AI-generated visual content presented as authentic, and on AI-written posts that simulate genuine user reviews. Brand official accounts (pinpai hao) have slightly different rules than individual user accounts, but even brand content is expected to reflect real products and real scenarios. Studio photography of actual products is compliant. Lifestyle photography showing real people wearing real garments is compliant. AI-generated “virtual try-on” images or AI-designed “mood board” posts presented without disclosure are not compliant.
The Consequences for Brands That Ignore the Policy
For brand accounts, violations result in content removal as a first step. Repeated violations lead to reach throttling, where the account’s posts are shown to a reduced audience. Severe or systematic violations can result in account suspension or permanent removal. For brands running paid advertising via Juguang, policy violations on the organic account can affect advertising eligibility. The risk is not just reputational: a suspended brand account on XHS means losing the verification status, follower base, and content history that took months to build.
How Fashion China Agency Builds Compliant XHS Content
At Fashion China Agency, our Xiaohongshu marketing team produces all brand content with human creators: real photography, real written posts in authentic Chinese consumer voice, and KOL collaborations with genuine XHS users. We do not use AI-generated imagery in client content on XHS, and our Chinese copywriting team produces text that reads as authentic because it is written by real people for a Chinese audience. This approach protects brand accounts from policy risk and produces content that actually performs: authentic-feeling content drives higher engagement on a platform built around peer recommendation. For beauty brands navigating the same restrictions on XHS, our partner Cosmetics China Agency runs compliant content programs for skincare and cosmetics brands. Contact our team via our contact page to discuss your brand’s XHS content strategy.
Sources: Xiaohongshu community guidelines update 2025 (xiaohongshu.com) | Reuters China AI content regulation overview 2025 (reuters.com)
Olivier Verot is co-founder and CEO of GMA (Gentlemen Marketing Agency) and Fashion China Agency. He manages XHS content programs for international fashion brands and monitors platform policy changes in China. Based in Shanghai since 2012. Connect on LinkedIn.