China · Beijing · Copyright

Beijing AI Text-to-Image Case

Evaluating prompt design, parameter adjustments, and the generation process, the court held that the disputed image reflected the user's intellectual investment and personalized expression, determining copyrightability and ownership on this basis.

Court or tribunal
Beijing Internet Court
Decision date
26 November 2023
Procedural status
First-instance judgment
Case number
(2023)京0491民初11279号

01

Factual background & dispute

  • The plaintiff generated an image using a generative artificial intelligence tool and published it on an online platform.
  • The defendant used the disputed image in an article and removed the original attribution watermark, prompting the plaintiff to file a lawsuit.
  • The court examined the prompts, parameter settings, modification process, publication records, and the alleged infringing use.

02

Core issues & judicial focus

  1. Whether an AI-generated image qualifies as a work protected under copyright law
  2. How human input during generation affects authorship determination
  3. What legal liabilities arise from reprinting content and removing attribution

03

Judicial finding & holding

  • The court held that the disputed image satisfied the requirement of originality.
  • Artificial intelligence models lack legal personality, making human input during the specific generation process a key basis for determining authorship.
  • The defendant was ordered to cease infringement, issue a formal apology, and pay damages.

04

Practical risk implications

01Maintain records of prompts, parameters, model versions, generation batches, and manual modifications.

02Verify material sources, model licenses, and similarity risks before commercializing generated content.

03Platforms and content teams should preserve attribution and publication metadata.

Court documentsFull judgments and orders1 document

PROJECT ENQUIRY

How does this precedent affect your product or litigation?

Consult with our practice team →