By Nila Kartika Wati 
A significant legal precedent is unfolding in Germany regarding the intersection of artificial intelligence and copyright law, with the Düsseldorf Higher Regional Court ruling that the transformation of a copyrighted photograph into an AI-generated comic-style image does not automatically constitute an infringement of the original photographer’s rights. This decision, handed down on April 2, 2026, in preliminary injunction proceedings, stems from a dispute involving a distinctive underwater photograph of a dog reaching for a red toy, and its subsequent AI-rendered adaptation. The ruling, while not a full trial on the merits, offers crucial insights into how German courts are beginning to grapple with the complex challenges posed by generative AI to established intellectual property frameworks, particularly in the realm of visual arts.
The Genesis of the Dispute: A Canine Underwater Portrait
The core of the Düsseldorf case revolves around a photographer specializing in underwater dog portraits. Her unique artistic style captures canines in dynamic aquatic environments, a niche that has garnered commercial success. The specific image at the heart of the litigation depicted a dog submerged, with its head and paws prominently visible as it extended towards a vibrant red toy. The photographer’s artistic accomplishment, as later noted by the court, lay in the precise interplay of sharpness and movement, achieved through tight framing and a shallow depth of field, which caused much of the dog’s body to recede into a gentle blur. This distinctive visual signature forms the basis of her protected work.
The legal challenge arose when a former cooperation partner, who operates a dog school, obtained this photograph. Instead of using the original, the partner uploaded the image into an AI software program, inputting prompts to generate a new, derivative image. The resulting AI-generated output, a comic-style rendition of the original scene, was subsequently published on the dog school’s website. The photographer, asserting a violation of her copyright, issued a cease and desist letter. When this demand went unheeded, she sought a preliminary injunction to prevent further use and distribution of the AI-generated image.
Chronology of the Düsseldorf Case and Legal Nuances
The dispute first reached a lower court, which ruled in favor of the defendant, classifying the AI-generated image as a permissible "free adaptation" (freie Benutzung) under German copyright law. This concept allows for the use of a protected work as inspiration for a new, independent work, provided the new work has sufficient creative distance from the original. However, the photographer appealed this decision, bringing the case before the Düsseldorf Higher Regional Court.
On April 2, 2026, the Higher Regional Court affirmed the lower court’s decision to deny the preliminary injunction, though it notably arrived at this conclusion via a different legal rationale. The appeals court explicitly disagreed with the lower court’s characterization of the AI image as a "free adaptation." Instead, the Higher Regional Court found no direct copyright infringement, emphasizing that the AI output had sufficiently transformed the protected elements of the original photograph.
Defining Originality and Transformation in German Law
To understand the court’s reasoning, it is essential to delve into the principles of German copyright law, known as Urheberrecht. Unlike common law systems that primarily focus on "copyright," Urheberrecht emphasizes the moral rights of the author (Urheber) and the inherent connection between the creator and their work. Central to this system is the concept of Schöpfungshöhe, or "threshold of originality." For a work to be protected by copyright, it must exhibit a sufficient level of individual intellectual creation. While photographs generally meet this threshold, the protection extends to the specific expression of an idea, not the idea itself. As legal analysts have frequently pointed out, the concept of "a dog reaching underwater for a toy" is, in itself, not eligible for copyright protection; it is a natural scene that a photographer might encounter rather than create.
The Düsseldorf court meticulously compared the original photograph with the AI-generated image. The original was characterized by its tight framing, the specific angle, the particular lighting conditions, and the shallow depth of field that blurred the dog’s body, leaving primarily its head and the toy in sharp focus. This interplay of sharpness and movement was deemed the unique artistic accomplishment. In stark contrast, the AI image presented the entire dog in a flat, cartoonish style, with exaggerated paws. Crucially, the court found that the AI output had transformed the very elements that constituted the photographer’s protected expression: the specific framing, the precise angle, the nuanced lighting, and the distinct sharpness. While the same dog and the overall setting were recognizable, this alone was deemed insufficient to prove direct copying. The court underscored that the idea of the scene was carried over, but the specific creative choices that elevated the photograph to a protected work were not replicated.
Broader Context: The LAION Case and a Developing Trend
This Düsseldorf ruling does not exist in isolation. It aligns with another significant German AI copyright battle involving renowned stock photographer Robert Kneschke. Kneschke sued LAION, a non-profit organization that compiles large datasets for AI training, after discovering his copyrighted images within the LAION-5B collection. In that case, the Higher Regional Court of Hamburg dismissed Kneschke’s appeal, affirming that the creation of the LAION-5B dataset did not infringe copyright.
The Kneschke case fundamentally addresses a different facet of the AI-copyright conundrum: the legality of using copyrighted material for training AI models. The Hamburg court’s decision suggested that merely ingesting images into a dataset for machine learning purposes, without direct reproduction or public display of the original works, does not constitute copyright infringement. This contrasts with the Düsseldorf case, which focuses on the copyright implications of the output generated by AI.
Read together, these two rulings reveal a nascent, yet consistent, judicial trend in Germany. German courts have, thus far, shown reluctance to expand existing copyright protections to encompass either the ingestion of images for AI training or AI-generated outputs that draw inspiration from a subject or concept without replicating the specific, protected creative choices of the original work. Both rulings are still subject to potential further legal challenges; the Kneschke case, notably, has been cleared for an appeal to Germany’s Federal Court of Justice (Bundesgerichtshof), indicating that the legal landscape remains fluid and subject to higher judicial review.
Legal and Industry Reactions: A Spectrum of Concerns and Optimism
The Düsseldorf decision has elicited varied reactions across the creative industries, legal community, and technology sector. For many photographers and artists, the ruling is a cause for significant concern. Associations representing visual artists have expressed apprehension that such interpretations could weaken copyright protections, making it increasingly difficult for creators to safeguard their unique styles and livelihoods. There is a palpable fear that if AI can freely "transform" elements without direct infringement, it could lead to a proliferation of derivative works that dilute the value of original creations. Creators often invest years in developing a recognizable aesthetic or a distinctive photographic approach, and the prospect of an AI tool being able to mimic or draw upon these elements without legal consequence is unsettling. They argue that existing copyright laws, conceived in an era before generative AI, are ill-equipped to address these novel challenges and may require legislative updates to provide adequate protection.
Conversely, proponents of AI innovation and developers within the technology sector view the ruling as a pragmatic and balanced approach. They emphasize that AI tools are designed to be creative engines, not mere copying machines. They argue that requiring AI outputs to be completely divorced from any inspiration would stifle technological progress and limit the potential for new forms of artistic expression. From this perspective, the Düsseldorf court’s focus on the transformation of protected elements, rather than the mere recognition of a common subject, supports the idea that AI is generating new works rather than infringing existing ones. This aligns with the broader principle that copyright protects expression, not abstract ideas or concepts.
Legal experts, while acknowledging the immediate implications, often highlight the nuanced nature of these preliminary decisions. They underscore that these are lower-court rulings in a single jurisdiction, not universally binding global precedents. However, they also concede that these cases expose significant "gaps and inconsistencies" in how current copyright law, drafted for a pre-AI world, grapples with artificial intelligence. The challenge lies in striking a delicate balance: fostering innovation in AI development while simultaneously ensuring that creators are fairly compensated and their intellectual property is adequately protected. This balancing act is not unique to Germany but is a global dilemma that policymakers and judiciaries worldwide are actively navigating.
Broader Implications and Future Outlook
The practical takeaway from the Düsseldorf ruling is not a blanket permission for AI to "freely copy" any creative work. Instead, it reiterates a fundamental principle of copyright law in Germany and across the European Union: protection is afforded to your specific expression, your unique artistic choices, rather than the underlying subject or concept you captured. This means that a rival photographer is legally permitted to photograph the same dog with the same toy underwater, provided their execution (framing, lighting, depth of field, etc.) constitutes an independent creative act. The Düsseldorf ruling now extends this principle to AI, suggesting that an AI can be prompted to generate an image based on the same motif, as long as it does not directly replicate the original photographer’s specific, protected elements.
For working photographers and other visual artists, this presents a narrower shield of protection than many might assume they possess. If a significant portion of a photographer’s business relies on a repeatable concept, a distinctive visual style, or a recognizable series, this ruling underscores the critical importance of robust legal strategies. This includes meticulously drafted contracts, clear licensing terms, and stringent control over who has access to original digital files. The Düsseldorf case itself serves as a cautionary tale: it began precisely because a former partner, with access to the original work, utilized it as input for AI generation.
Looking ahead, these rulings underscore the urgent need for clarity and, potentially, legislative reform in the domain of AI and copyright. The European Union’s Copyright Directive (Directive 2019/790) includes provisions related to text and data mining (TDM), aiming to facilitate AI training while balancing creators’ rights. However, the interpretation and application of such directives, particularly concerning AI-generated outputs, remain subject to national court decisions and future legislative actions. The upcoming appeal in the Kneschke case to Germany’s Federal Court of Justice could be a pivotal moment, potentially setting a higher judicial standard for how AI’s interaction with copyright is understood.
The future landscape of intellectual property in the age of AI will likely involve ongoing debates about what constitutes "originality" when algorithms are involved, how to define "transformation" versus "derivative work," and whether entirely new legal frameworks are needed to address machine-generated content. These German decisions mark an important early chapter in this global conversation, exposing the existing law’s limitations and signaling that the line between inspiration and infringement, in the context of AI, is still very much in flux and subject to redefinition by future rulings.
