On 8 August 2025, the Hangzhou Intermediate People’s Court delivered judgment in Zhejiang’s first unfair-competition case involving the commercial application of AI technology. The plaintiff was a technology company operating a social e-commerce platform; the defendants were two internet companies that developed an AI writing tool. The court found that the defendants’ bulk copy-generation service constituted unfair competition and ordered them to pay RMB 100,000 jointly.
The core issue
The defendants’ AI writing tool was designed specifically for the “seed planting” scenario on a social e-commerce platform and offered bulk copy generation. The plaintiff argued that generating marketing content in bulk for a specific platform scenario disrupted the platform’s normal operating order.
The court’s holding rests on a clean dividing line: the technology itself is neutral, but its application is subject to constraints. Generative AI services delivered at the application layer for a specific scenario should respect the rules of that scenario.
Why the distinction matters
If the analysis stopped at “AI is a tool”, bulk copy generation would be no different from typing on a keyboard. The court instead moved the level of analysis down from technology to application scenario: the same technology is neutral when used as a general writing assistant, but may constitute unfair competition when deployed specifically to circumvent a platform’s content-ecosystem rules.
The applicable provision is the principle-based Article 2 of the Anti-Unfair Competition Law. That means even absent a specific prohibition, a court can still make a finding on the basis of good faith and business ethics.
Echoes of the AI model protection cases
In the same period, the Supreme People’s Court Intellectual Property Tribunal published a typical case confirming that the parameters and structure of an AI model formed through data training and tuning can constitute a legally protected competitive interest. See the Supreme Court AI model parameters case.
The two cases protect the same interest from different directions: the former protects the model itself (the supply side), while this case regulates how a model is applied (the application side). Together they show an emerging two-way framework of “model protection plus application constraint” in AI competition law. See the first AI model infringement case.
The Hangzhou courts’ approach
According to reports, Hangzhou courts have accepted a number of new-type cases involving AI and have issued judicial opinions on supporting an AI innovation hub, setting out adjudicative philosophies for several categories of typical disputes.
This shows that in an AI industry cluster, courts are supplying rules case by case, filling the gap left by slow legislation. See the Shanghai AI prompt copyright case and the Hangzhou AI hallucination tort case.
Where businesses can start
First, assess the rules of the target scenario of an AI product. Tools aimed specifically at one platform should be checked against that platform’s content and competition rules.
Second, avoid positioning a product as a tool for circumventing rules. The marketing language itself can become adverse evidence.
Third, retain evidence of investment in training and tuning to preserve room to assert your own competitive interest.
The most important lesson is that the compliance risk of an AI product often sits not at the model layer but at the level of product design and scenario choice.
This article is based on public reports on the unfair-competition case concerning the commercial application of AI technology decided by the Hangzhou Intermediate People’s Court and is for reference only; it does not constitute legal advice.