A screenshot shows an AI writing tool offering one-click generation services for recommendation-style posts and other online content. Photo: Screenshot from CCTV News

The Hangzhou Intermediate People’s Court in East China’s Zhejiang Province recently disclosed a typical case involving the mass generation of “social recommendation posts” for a social media platform using AI, marking the first case of its kind in China involving AI-written recommendation posts, state broadcaster CCTV News reported on Tuesday.  

The case comes as a wave of AI writing tools promising to “generate viral recommendation posts with one click” has emerged online. At extremely low cost, these tools can mass-produce fake consumer experiences and product endorsements.

The plaintiff, the operator of a well-known social media platform whose rules and community guidelines require users to share real experiences, sued two companies, identified as Company B and Company C, over an AI writing tool they jointly operated.

According to CCTV, the tool offered one-click generation of recommendation-style posts, travel guides and other content tailored to the style of the platform, while encouraging users to publish the generated copy there.

The plaintiff argued that such conduct damaged the platform’s ecosystem of authentic recommendation content, and asked the court to order the defendants to stop copyright infringement and unfair competition.

The report, citing Zhang Zongbin, a judicial assistant in the case, as saying that the tool was available as a web version, an app and a WeChat mini program. Its popular modules directly displayed services such as platform-style recommendation posts, travel guides and note headlines.

The modules also used names associated with the platform and promoted services such as generating “sharing copy in line with the platform’s tone based on product features.” With one click, users could produce a post of several hundred words.

The central question before the court was whether the targeted provision of generative AI services for the social media platform constituted unfair competition.

In its review, the court first made clear that the authentic content ecosystem the plaintiff had built through long-term operations was a competitive interest protected by law. Given the nature of generative AI services, the court found that the defendants had failed to fulfill a reasonable duty of care in both the way the tool was operated and the purpose it served.

The court ordered Company B and Company C to pay the plaintiff a total of 100,000 yuan ($14,716.7) in compensation for economic losses and reasonable expenses, according to the report.

Liu Dingding, an internet industry analyst, said the ruling carries landmark significance, as it could be seen as an important step by the social media platform in tackling AI-generated fake content.

According to Liu, the core issue in the case is not AI-generated content itself, but the falsified nature of the content and the deliberate traffic-diversion practices behind it.

“There is nothing inherently wrong with using AI to assist with copywriting. The problem is that these tools are generating fabricated consumer experiences and fake recommendation content, which is already a violation in itself,” he said.

The ruling not only addressed the plaintiff’s claims, but also clarified the boundaries for providers of generative AI services, according to the CCTV report.

The ruling came amid the wider proliferation of low-cost online courses, cracked software and tutorials related to “AI automated account operation” and “matrix account building” on short-video and second-hand trading platforms.

Under product recommendation functions, users can generate recommendation-style posts of several hundred words within seconds simply by entering details such as a product name and brand. These posts, often disguised as genuine personal reviews or firsthand experiences, can be produced entirely by AI without any actual travel experience or product use, with the content fully fabricated by the system.

Liu said such low-cost, low-barrier content generation makes it increasingly difficult for consumers to distinguish genuine reviews from AI-generated fake endorsements, while also putting pressure on original content creators and disrupting the platform’s content ecosystem.

The Cyberspace Administration of China recently launched a special campaign targeting AI-related online misconduct, including the use of AI “hosting” technologies to operate accounts, simulate real users in batch account registration and management, and generate and post large volumes of low-quality, repetitive content.

The case also signals that China is further clarifying the legal boundaries for the commercial use of generative AI, Liu said, adding that “China supports AI innovation and industrial development, but AI is not beyond legal regulation.” 

But addressing disorder in AI applications will also require content platforms, AI model developers and generative AI service providers to work together and build a full-chain governance framework, he added.