The New Paradigm of Originality in Copyright Law

View on LinkedIn

Originality has traditionally been the cornerstone of copyright law, the fundamental requirement that determines whether a work deserves legal protection. In Mexico, this principle is enshrined in Article 3 of the Federal Copyright Law, which establishes that "the works protected by this Law are those of original creation capable of being disclosed or reproduced in any form or medium." However, the arrival of artificial intelligence has prompted a profound reexamination of what it truly means to be "original" in the twenty-first century, especially when machines can generate creative content autonomously or semi-autonomously.

To understand the complexity of this concept in today's context, we must begin with its multiple dimensions. The word "original" carries two principal meanings according to the Royal Spanish Academy: it can refer both to the origin or starting point of something, and to that which is novel or singular. This duality is reflected in copyright law, where we speak both of the "original" as the primary work from which other creations may derive, and of "originality" as the distinctive creative quality that separates a protectable work from the commonplace or trivial.

The current legal framework establishes certain basic parameters. Jurisprudential criterion 2027524 recognizes protection for both original and derivative works, provided they display their own characteristics of originality. The Berne Convention, for its part, extends this protection to translations, adaptations, and musical arrangements, treating them as original works so long as they respect the rights of the author of the underlying work.

In practice, however, the application of the originality concept reveals numerous paradoxes and unwritten criteria that vary according to the type of work and its context. For example, a literary translation generally receives protection as an original derivative work, even though its primary objective is to faithfully reproduce the source text in another language. By contrast, certain creations that might seem more "original" in the common sense of the word — such as an innovative functional design or a clever advertising slogan — may fail to reach the threshold of protection because they are deemed too brief or too functional.

This selective flexibility of the originality concept manifests in seemingly contradictory ways. While some works are excluded for their brevity or simplicity, others receive protection despite following highly standardized patterns. The formats of legal documents or contracts, for instance, are rarely considered original, whereas musical arrangements that follow established conventions may indeed receive protection. These disparities suggest that originality in copyright law is not a uniform concept, but rather a flexible criterion that adapts — sometimes capriciously — according to the type of work, its context, and the established practices within each creative field.

In this context, it is understandable and coherent that works generated in a fully automated manner, without any form of human creative intervention or effort, should fail to reach the originality threshold required for copyright protection. This requirement of human participation in the creative process is not a mere legal whim, but a fundamental principle that distinguishes genuine artistic creation from the mere mechanical generation of content.

The debate over the fate of these purely automated works raises fascinating questions that transcend the issue of originality. While some voices suggest that such content should pass directly into the public domain — arguing that the absence of a human author implies the absence of exclusive rights — this discussion of economic ownership is a separate debate that warrants its own detailed analysis. What is clear is that the absence of originality, understood as the lack of human creative intervention, constitutes a fundamental barrier to copyright protection, regardless of any discussion about the eventual ownership or management of such content.

In doctrinal analysis, the concept of the "author's imprint" has been fundamental to understanding originality. This notion — referring to the personal and unique mark that the creator leaves on a work — is widely used in legal scholarship and case law to determine copyright protection. It is crucial to note, however, that this concept is not codified in any statute or international treaty; rather, it is a doctrinal construct that helps interpret and apply the originality requirement.

This absence of a statutory definition of the "author's imprint" takes on particular relevance in the current context. When the United Kingdom recognizes rights for computer-generated works, a fundamental question arises: are the other member states of the Berne Convention obligated to recognize these rights under the national treatment principle? This question poses a significant challenge to the international copyright system, as it could create tension between the different national approaches to originality and authorship in the age of AI.

The "Zarya of the Dawn" case represents an inflection point in this discussion. Resolved by the United States Copyright Office in February 2023, this case offers a thorough and meticulous analysis of how originality should be understood when artificial intelligence participates in the creative process.

The case began when Kristina Kashtanova applied for copyright registration of a graphic novel created using Midjourney, an AI tool that generates images from textual descriptions. To understand the complexity of the case, it is essential to understand how this tool works. As the Copyright Office explains in its ruling, Midjourney does not interpret instructions (prompts) as specific commands to produce a particular result. Instead, the tool converts words and phrases into "tokens" that it compares against its training data. The process begins with a field of random visual noise, similar to television static, and uses algorithms to gradually refine that noise into recognizable images, generating multiple variations that the user may select and refine.

Kashtanova's creative process was extensive and meticulous, as her attorney Van Lindberg explained. Each image required hours of work, and each page at least a full day. The work grew out of a deeply personal experience — the loss of her grandmother (Raya) and her best friend — which was reflected in every aspect of the creation.

The arguments presented by Kashtanova and her attorney were substantial. First, they emphasized the integrity of the creative process, highlighting that each image was the result of an extensive and deliberate process requiring constant interaction and creative decision-making. They also argued for the conceptual depth of the work, pointing to how each visual and narrative element was connected to Kashtanova's personal history.

A central argument was the characterization of the work as a compilable work protectable under Section 101 of the Copyright Act, highlighting the selection, coordination, and arrangement of elements as protectable creative acts. They drew an analogy to photography, comparing the use of Midjourney to the photographic process and equating decisions about composition, selection, and framing with the creative decisions of a photographer.

Particularly noteworthy was their argument treating the prompts as creative works in their own right, comparing them to poems that paint pictures with words and contending that the generated images were representations of this initial creativity. They also described a refinement process that involved hundreds of versions of each image, with constant selection, combination, and modification of elements.

The Copyright Office conducted a careful analysis of the case and reached nuanced conclusions. It recognized full protection for the text written by Kashtanova, determining that it was a direct product of human creativity and met the traditional requirements of originality. It also recognized protection for the work as a compilation, valuing the selection, coordination, and arrangement of elements as decisions reflecting protectable creativity. However, it determined that the individual images were not protectable by copyright, emphasizing the lack of direct control over their generation and the unpredictability of the final result.

The Office specifically rejected the "sweat of the brow" argument, establishing that the effort invested, however significant, does not constitute a sufficient basis for protection. It also elaborated an illuminating analogy: giving instructions to Midjourney was similar to giving instructions to an artist, and instructions alone do not constitute authorship of the resulting work. Regarding the modifications made in Photoshop, the Office specifically analyzed two images with manual edits but determined that the changes were "too minor and imperceptible" to constitute protectable originality.

In contrast to this approach, the United Kingdom has developed a significantly different one. The Copyright, Designs and Patents Act 1988 specifically recognizes "computer-generated works" as a special category of protectable creation. The Act defines these works as those created without a direct human author and establishes a clear legal framework for their protection.

The British model takes a pragmatic approach by designating as author "the person by whom the arrangements necessary for the creation of the work are undertaken," thereby resolving the practical problem of rights attribution. It also establishes a limited protection term of 50 years, differentiating these works from traditional creations while maintaining incentives for their production. The originality requirement is preserved but adapted to the technological context, recognizing new forms of creative expression.

This contrast between the American and British approaches reveals different ways of adapting copyright law to the digital age. Whereas the United States emphasizes direct control over the result and maintains traditional categories of protection, the United Kingdom recognizes broader forms of creative contribution and creates new specific categories for these works.

The combined experience of the Zarya case and the British model suggests that the concept of originality is evolving along multiple dimensions. On the technical side, we must reconsider the relationship between technical control and authorship, the role of automated tools, and the limits of human intervention. On the creative side, we need to understand the changing nature of creative contribution and the new forms of originality that are emerging. On the legal side, we face the challenge of adapting existing legal frameworks and creating new categories of protection that maintain a balance between innovation and protection.

The current challenge is not simply to determine whether AI-created works are original, but to develop a conceptual and legal framework that recognizes new forms of creativity, effectively protects the rights of all parties involved, and promotes innovation while upholding the fundamental principles of copyright law. These include the protection of creativity, the recognition of authorship, and the promotion of cultural progress.

The debate over originality in the age of AI will continue to evolve. The Zarya case and the British model are only the first steps along a path that will continue to develop alongside the technology and our understanding of human creativity. The key will be to find a balance that promotes innovation while protecting the rights of all those involved in the creative process.

This debate extends well beyond AI-generated images. In the field of software development, for example, tools such as GitHub Copilot and Amazon CodeWhisperer are transforming the way software is built. What happens to code generated or suggested by these tools? The line between assistance and authorship becomes especially blurred when we consider that code must be not only functional but also original to merit copyright protection.

In the musical field, AI-assisted composition and arrangement tools raise similar questions. Software such as OpenAI's MuseNet or Google's Magenta can generate complete musical compositions or suggest arrangements based on existing patterns. When a musician uses these tools to create a new version of a work, where does technical assistance end and where does protectable creative expression begin?

The creation and editing of video presents another set of challenges. With AI tools capable of generating, editing, and transforming audiovisual content, the determination of originality becomes increasingly complex. Deepfake and video synthesis technologies can produce content that appears entirely original but derives from existing works in ways that are not readily apparent.

In the publishing world, the use of AI for editing, proofreading, and even content generation is transforming traditional creative processes. When an author uses AI tools to refine their prose, suggest narrative structures, or even generate sections of content, how do we evaluate the originality of the final result?

These questions suggest that we need a more sophisticated framework for evaluating originality in the digital age. That framework should:

First, recognize different levels of AI intervention, distinguishing between tools that assist the creative process and those that generate content autonomously. The key may lie in assessing the degree of human control and direction over the final result.

Second, consider the transformative nature of AI use. When technology is used to transform, combine, or reinterpret existing works, we need clear criteria to determine when the result constitutes a new original work.

Third, adapt our understanding of the "author's imprint" to encompass new forms of creative expression. The author's creative mark may manifest not only in the direct creation of content, but also in the selection, curation, and organization of AI-generated elements.

Fourth, develop specific standards for different creative fields. The standard of originality for software code may require criteria different from those applied to music or the visual arts.

The future of copyright law will depend on our ability to balance the protection of human creativity with the recognition of new forms of expression enabled by technology. As AI continues to evolve, our concept of originality must evolve with it, always keeping as its guiding principle the promotion and protection of human creativity in all its manifestations.