Global Courts Grapple with AI-Generated Content and Copyright: Key Cases from China, UK, and Implications for Türkiye

Artificial intelligence (AI) technologies have rapidly integrated into daily life, generating new outputs by processing inputs such as text and images. This widespread adoption has intensified legal debates over whether AI-generated outputs qualify as "works" under copyright law and whether they infringe third-party rights. These discussions have moved beyond theory into concrete judicial decisions across multiple jurisdictions.

Central to these debates is the concept of "authorship," a cornerstone of copyright law. Courts worldwide have grappled with whether AI-generated content can be attributed to human creators for copyright protection. A notable example is the 19 March 2025 ruling by the Zhangjiagang People’s Court in Jiangsu Province, China. In this case, designer Feng Runjuan used the AI image-generation platform Midjourney to create an image based on the prompt: "children’s chair with a jelly texture, shape of cute pink butterfly, glass texture, light background." After Feng declined a production offer from Kuashi Plastic, the company allegedly manufactured and sold chair models closely resembling Feng’s AI-generated design.

Feng filed suit alleging copyright infringement and unfair competition. The defendants contended that the plaintiff’s AI prompts did not constitute protectable intellectual creations and that similar images predated Feng’s design. The court noted the inherent unpredictability of AI outputs, as identical prompts may yield different results. Ultimately, the court held that copyright protection does not arise when content is predominantly AI-generated. It emphasized that human authorship requires intentional, decisive creative choices—such as individualized perspective, color, and composition—clearly demonstrated by the claimant. Feng’s failure to meet these criteria led to dismissal of both claims.

Similarly, the Beijing Internet Court addressed a dispute involving an image described as a "crystal cat necklace" allegedly created with AI and used unlawfully by the defendant for promotion. The court outlined cumulative criteria for determining whether an output qualifies as a "work": (i) belonging to literature, art, or science; (ii) originality; (iii) expression in a specific form; and (iv) intellectual creation. Regarding originality, the plaintiff bore the burden of proving the creative process with concrete evidence. On 16 September 2025, the court dismissed the claims due to insufficient evidence of originality.

These Chinese rulings illustrate a prevailing global approach: copyright protection for AI-generated outputs hinges on demonstrable human creativity and control over the creative process. Courts assess the extent to which human inputs decisively shape the final output.

Beyond authorship, the use of copyrighted works in AI training datasets remains a contentious issue. Lawsuits worldwide challenge whether generative AI models unlawfully reproduce protected works during training. In a prominent case, Getty Images sued Stability AI in the High Court of England and Wales, alleging unauthorized use of Getty’s images in AI training, infringing reproduction rights. The court’s 4 November 2025 decision held that mere exposure of an AI model to copyrighted works during training does not constitute infringement, as the model does not store or contain infringing copies. Consequently, the copyright claims were dismissed.

While this ruling clarifies that AI training on copyrighted materials is not automatically infringing, it does not resolve ongoing debates about the scope of such use or the nature of AI-generated outputs that may substantially replicate protected works. Different jurisdictions may apply varying standards, particularly when outputs closely resemble pre-existing works.

Under Turkish law, Article 22 of the Law on Intellectual and Artistic Works explicitly regulates reproduction rights, granting authors exclusive rights to reproduce their works by any means, directly or indirectly, temporarily or permanently. This framework suggests that copying involved in AI training could be classified as temporary or indirect reproduction, potentially differing from the UK court’s approach.

To date, Türkiye has not issued court decisions directly addressing AI and copyright issues. However, given the rapid proliferation of AI technologies, significant legal developments in this area are anticipated in the near future.

In conclusion, as of 2026, the intersection of artificial intelligence and copyright law remains dynamic and evolving. Courts worldwide emphasize the necessity of human creative input for copyright protection of AI-generated works, while the legality of AI training on copyrighted materials continues to provoke complex legal questions. Navigating this landscape requires careful management of legal risks and vigilant monitoring of emerging regulations to balance technological innovation with the protection of intellectual property rights.