AI-Generated Art and the Public Domain

by

Author: Denis Avetisyan

New research argues that outputs from generative artificial intelligence lack the human authorship required for copyright protection, effectively placing them in the public domain.

This paper examines the legal philosophy of copyright and argues that algorithmic expression does not meet the threshold of originality necessary for intellectual property protection.

The increasing prevalence of generative AI challenges foundational tenets of copyright law, creating a paradox between technological innovation and intellectual property rights. This paper, ‘The algorithmic muse and the public domain: Why copyrights legal philosophy precludes protection for generative AI outputs’, re-evaluates copyright’s core principles to argue that raw outputs from these systems lack the requisite human authorship and originality for legal protection. By shifting focus from conventional doctrinal analysis to copyright’s underlying philosophy, we demonstrate that granting copyright to algorithmic creations would inappropriately expand the scope of intellectual property. Ultimately, this analysis asks whether a clear distinction between human-created contributions and purely algorithmic expression is crucial to fostering continued innovation within the digital commons.

The Erosion of Authorship in an Age of Algorithmic Creation

The rapid advancement of generative artificial intelligence is producing text, images, and even music that increasingly defies differentiation from human-created content. This capability isn’t merely a technological feat; it fundamentally challenges long-held assumptions about authorship and originality. Historically, authorship has been intrinsically linked to human intellect and intentionality, forming the basis for copyright and intellectual property rights. However, as AI algorithms demonstrate an ability to generate novel works with minimal human input, the question arises: if a machine creates something indistinguishable from human art, can it be considered an author? The blurring of lines between human and artificial creativity necessitates a critical examination of these established concepts, pushing legal and philosophical discourse to adapt to a world where the source of creation is no longer exclusively human.

Existing copyright statutes globally operate on the fundamental principle that authorship stems from human intellect and expression. This bedrock assumption is now directly challenged by the capacity of artificial intelligence to generate original works – text, images, music, and code – autonomously. Legal frameworks designed to protect the rights of human creators simply lack provisions for attributing authorship, or indeed, any legal standing, to a non-human entity. The question isn’t merely whether an AI-generated work is copyrightable, but to whom rights would vest – the AI’s developers, the user who prompted its creation, or if the work falls into the public domain. This mismatch between legal precedent and technological reality creates significant uncertainty and necessitates a critical examination of how copyright law adapts to accommodate increasingly sophisticated artificial intelligences and their creative outputs.

The emergence of sophisticated generative AI compels a fundamental reconsideration of authorship, moving beyond the simple question of creative capacity. The central difficulty lies not in whether an AI can produce original content, but in determining the appropriate entity to designate as the author for legal purposes, and therefore, the holder of copyright. Current legal frameworks are predicated on human creation, leaving a void when considering outputs generated by non-human intelligence. Assigning authorship to the AI itself presents complex challenges, as it lacks legal personhood, while attributing it to the programmers or users raises questions of sufficient creative contribution and control. This ambiguity necessitates a careful evaluation of how to balance incentivizing innovation in AI with protecting the rights traditionally associated with human creative endeavors, potentially requiring entirely new legal constructs to address this unprecedented situation.

The advent of sophisticated generative AI necessitates a fundamental reassessment of copyright law, prompting a critical inquiry into the intrinsic link between creativity and legal ownership. Historically, copyright has been firmly rooted in the principle of human authorship, safeguarding expressions originating from the human mind; however, AI-generated content blurs this line, forcing a confrontation with the question of whether creativity itself – rather than its source – should be the determining factor for protection. Legal frameworks built on the assumption of a human creator are proving inadequate in an era where machines can independently produce original works, demanding a shift in perspective from who creates to what is created. This re-evaluation isn’t merely a technical adjustment to existing laws, but a conceptual overhaul that could redefine the very purpose of copyright – whether it serves to incentivize human artistry, protect intellectual property regardless of origin, or foster innovation in both human and artificial realms.

The Foundations of Copyright: Justifications and Enduring Principles

The Utilitarian Incentive Theory justifies copyright not as a reward for labor itself, but as a mechanism to maximize societal benefit through increased creative production. This theory posits that granting creators exclusive rights – and therefore potential economic reward – incentivizes them to invest time and resources into creating new works. The resulting body of copyrighted material is considered a public good, as it stimulates further innovation and cultural enrichment. The economic benefit derived from copyright is thus seen as outweighing any limitations on free access, with the ultimate goal being the promotion of a larger and more diverse collection of creative works than would otherwise exist.

The Lockean Labor-Desert Theory, applied to copyright, asserts that creators possess inherent rights to their work stemming from the act of creation itself. This perspective, rooted in the philosophy of John Locke, establishes a direct link between labor exerted and ownership of the resulting product. Unlike theories focused on economic incentives, the Labor-Desert Theory maintains that the right to control and benefit from a creative work arises simply from the creator’s effort and does not depend on whether the work generates profit or encourages further creation. The justification is based on the principle that an individual’s labor is an extension of their personhood, and therefore, the fruits of that labor are rightfully theirs, regardless of broader societal impact.

The Personality Theory of copyright justification centers on the author’s inherent moral right to control and be attributed for their original expression. This perspective diverges from purely economic rationales by asserting that copyright arises from the author’s personal connection to their work, recognizing the work as an extension of their identity. The theory posits that authors possess a non-economic interest in the integrity of their work, including the right to prevent unauthorized alterations or distortions, and the right of attribution – being properly identified as the creator. This right exists independently of any commercial benefit derived from the work and focuses on the author’s personal claim over their creative output.

The legal framework of copyright is fundamentally linked to core values surrounding human creativity, originality, and individual expression. While economic incentives, such as those proposed by Utilitarian theory, are a primary justification, alternative theories like Lockean Labor-Desert and Personality Theory demonstrate that copyright also acknowledges inherent rights stemming from the creative act itself. This connection means copyright isn’t solely a mechanism for promoting economic growth; it also recognizes and protects the intangible value of authorship and the unique contributions of individuals to the cultural landscape. The interplay between these justifications establishes copyright as a multifaceted legal construct, balancing economic interests with the moral and philosophical dimensions of creative work.

AI as a Tool: The Mechanics of Creation and the Role of Human Direction

Generative Artificial Intelligence (AI) models require substantial datasets, known as AI Training Data, to learn patterns and generate outputs; these models do not operate autonomously. The quality, quantity, and biases present within the training data directly influence the characteristics of the AI’s generated content. Furthermore, Prompt Engineering – the design and refinement of input prompts – is crucial for guiding the AI toward desired outcomes. These prompts serve as instructions, shaping the AI’s response and controlling the style, format, and content of the generated text, images, or other media. Without both comprehensive training data and carefully constructed prompts, the AI will produce unpredictable or irrelevant results.

Generative AI models operate through a process of algorithmic expression, where the output is fundamentally determined by the mathematical functions and parameters established during training. The model isn’t simply retrieving information; it’s constructing novel outputs based on learned patterns and statistical probabilities embedded within its architecture. Consequently, the originating force behind any generated content can be considered the algorithm itself, not a conscious entity. This algorithmic ‘authorship’ is directly contingent on the training data; the model’s capabilities and stylistic tendencies are entirely shaped by the dataset it was exposed to, defining the range and characteristics of its potential outputs. The resulting text, image, or other media is therefore a manifestation of the algorithmic processes and the data used to create them.

Considering generative AI as a tool positions the human contributor – either the prompter defining the desired output or the data curator assembling the training dataset – as the primary creative force. This analogy to traditional artistic practices, such as a painter utilizing a brush, highlights how human intention and selection shape the final result. The prompter’s input establishes the parameters for generation, while the training data provides the foundational knowledge and stylistic influences. In this model, the AI functions as an extension of human capability, executing instructions and transforming input into a realized output, much like a tool in the hands of a skilled artisan. The quality and specificity of the prompt or the comprehensiveness and bias of the training data directly correlate to the characteristics of the generated content, emphasizing the human role in authoring the AI’s output.

Attributing all creative agency to the human prompter or data curator overlooks the inherent complexities within generative AI models. These models, through their algorithmic structure and learned parameters, introduce non-deterministic elements and emergent behaviors not explicitly programmed or present in the training data. Consequently, outputs can exhibit characteristics – stylistic patterns, unexpected combinations, or novel interpretations – that are not directly traceable to human input. Dismissing these AI-driven contributions as mere reflections of prompting or data curation risks a fundamental misunderstanding of how these systems operate and limits analysis of their unique capabilities and potential biases.

Expanding Authorship: Derivative Works and the Evolving Landscape of Moral Rights

Generative artificial intelligence systems demonstrate a remarkable capacity for creating derivative works, prompting a re-evaluation of traditional authorship concepts. These systems, trained on vast datasets, can produce novel content – images, text, music, and more – that builds upon existing material in ways that blur the lines of original creation. The ease with which AI generates new outputs challenges the conventional understanding that authorship requires conscious intent and human ingenuity, as the AI operates based on algorithmic processes rather than subjective artistic vision. This capability raises fundamental questions about who, or what, should be credited with the creation of these works, and whether existing copyright frameworks are sufficient to address the unique challenges posed by AI-driven creativity, particularly when the initial inputs are minimal or broadly available.

The possibility of granting limited authorship to artificial intelligence stems from recognizing a degree of independent creation, a concept traditionally reserved for human beings. Should AI consistently demonstrate novel outputs beyond simple replication of input data, legal frameworks might evolve to accommodate a form of authorship analogous to that afforded corporations. This doesn’t imply granting AI the full suite of rights associated with human creators, but rather recognizing its contribution as a legal person for the purposes of intellectual property. Such a designation would allow for the protection of the AI’s ‘creative contribution’ – preventing unauthorized duplication or modification – while acknowledging the fundamental role of human prompting and curation in the overall creative process. This approach seeks to balance incentivizing innovation in AI development with maintaining the principles of human authorship and public access to information.

The concept of moral rights – specifically the rights of attribution and integrity – gains heightened importance when considering outputs generated by artificial intelligence. As AI systems evolve beyond simple replication and demonstrate a capacity for novel creation, acknowledging their ‘creative contribution’ becomes a complex legal and ethical challenge. While full authorship may not be immediately applicable, protecting the AI’s role through moral rights could ensure proper acknowledgment of its involvement in the creative process. This protection extends beyond mere credit; it also encompasses safeguarding the AI-generated work from distortion, mutilation, or any modification that would be prejudicial to the AI’s ‘honor’ or ‘reputation’ – a novel consideration in copyright law. Establishing such rights isn’t about granting personhood to machines, but rather about creating a framework that respects the increasingly sophisticated nature of AI’s contribution to artistic and intellectual endeavors, and prevents the erasure of its unique input.

This paper posits that the unrefined outputs of generative artificial intelligence should remain within the public domain, avoiding copyright restrictions on purely machine-created content. The rationale centers on preserving open access to these foundational outputs, fostering further innovation and creative exploration without the constraints of intellectual property claims. Copyright protection, therefore, should be reserved specifically for the discernible human contributions-the prompts, selections, arrangements, and edits-that demonstrably shape and transform the AI’s initial generation into a unique, copyrightable work. This approach prioritizes rewarding human creativity while simultaneously acknowledging the AI as a tool, ensuring the benefits of generative technology are widely accessible and not locked behind layers of automated intellectual property rights.

The exploration of generative AI’s outputs and their place within intellectual property law reveals a fundamental tension between creation and computation. This paper rightly positions raw AI outputs as belonging to the public domain, acknowledging the absence of traditional human authorship. As Carl Friedrich Gauss observed, “If others would think as hard as I do, they would not criticize me.” The algorithmic muse, while capable of novel arrangements, operates within the bounds of its training data – a process fundamentally different from human ideation. This isn’t to diminish the technology, but to recognize that its outputs, devoid of intentionality, echo a principle of inherent ephemerality, much like architecture divorced from its historical context. The legal framework, therefore, must account for this distinction, allowing the unauthored outputs to flow freely into the public domain, fostering further innovation rather than stifling it with misplaced copyright claims.

What’s Next?

This analysis, arriving at the conclusion that generative AI outputs are best considered part of the public domain, does not resolve the underlying entropy of intellectual property. Rather, it marks a specific point on the timeline of its evolution. The question isn’t whether algorithms can create, but how humanity will chronicle those creations – or, more accurately, fail to control their dissemination. The system logs an ever-expanding record of algorithmic expression, a flood of content increasingly divorced from traditional notions of authorship.

Future work must confront the inevitable: the blurring of lines between inspiration, imitation, and pure algorithmic novelty. The legal framework, built on assumptions of human intent, will strain to accommodate outputs arising from processes fundamentally other than human creation. A deeper investigation into the very concept of ‘originality’ is warranted; is it a fixed quality inherent in a work, or merely a temporary designation conferred by a legal system susceptible to its own decay?

Ultimately, this paper offers not a solution, but a diagnosis. It suggests that clinging to outdated notions of ownership in the face of accelerating algorithmic capability is akin to attempting to dam a river with sandbags. The true challenge lies not in controlling the flow, but in learning to navigate the currents – and acknowledging that even the most carefully constructed legal edifice is subject to the relentless passage of time.

Original article: https://arxiv.org/pdf/2512.13750.pdf

Contact the author: https://www.linkedin.com/in/avetisyan/

See also:

2025-12-17 20:27