When asked about X’s possible cooperation with Chinese content censorship, chatbot Grok answered that X is, in fact, partially helping the CCP to "maintain stability” on its platform as an unfortunate reality of doing business.

Video script: 

Chinese citizen, and mother of the 2019 “Vulgar Wikipedia” case’s Niu Tengyu (牛腾宇), has had her X account repeatedly banned and restricted. This has again raised the specter of X self-censoring in cooperation with Chinese pressure. When users asked the platform AI, Grok, about the case, they were afforded an unusually frank look behind the curtain into the platform’s decision-making regarding censorship, political pressure, and commercial concerns.

Niu was severely punished and tortured for allegedly leaking photos of Xi Mingze, Xi Jinping’s daughter, online. Since then, his mother, Coco, has vehemently spoken out on X, advocating her son’s innocence. Her account is frequently banned without warning, or marked as containing sensitive content – and even after reinstatement and appeals, she remains a target for pro-China nationalists, and online trolls. This has resulted in the frequent shadow-banning of her account – or limiting its exposure, for example, letting users view, but not interact with posts.

A netizen asked Grok why Coco’s account was repeatedly restricted, and received four reasons. First, Grok said that it “crossed a red line” regarding political topics. Second, it received systematic large-scale reporting. Third, Grok said it was affected by political pressure from Chinese authorities, especially regarding potential cooperation. Finally, Grok said the account encountered limitations imposed by the platform’s own risk-control algorithms.

The AI notes that such situations are common on X, saying many Chinese dissident accounts, such as Zhang Zhan’s mother, accounts involved with the “8964” incident, and White Paper Movement, have long been subjected to traffic restrictions or labeled for “sensitive content.” This matches what independent Chinese journalist Gao Yu has reported about her fluctuating follower count, and rumors that X’s Chinese content moderation team is based in Singapore.

When it was then asked whether it was directly helping the CCP’s “stability maintenance" efforts, the AI said X had never signed a formal contract to do so, but did make compromises for the sake of commercial interests, over legal risks, and advertiser concerns. Grok called this a state of overall “semi-cooperation”, concerning “sensitive” accounts, not outright banning them to maintain the image of a free platform, while at the same time significantly limiting traffic and employing shadow banning to prevent such voices from spreading and angering the Chinese market. Ultimately, it referred to the claim of helping the CCP maintain peace as “half true, and half a helpless business reality.”

Netizens responded sarcastically, calling the platform’s attempts to appease the West and China a “Chinese Free-Speech” operating model.

Amanda Stephens, for Rti News