COPYRIGHT IMPLICATIONS OF AI-GENERATED WORKS BY: GUNJAN RATHORE

“COPYRIGHT IMPLICATIONS OF AI-GENERATED WORKS”
 
AUTHORED BY: GUNJAN RATHORE
[B.A LL.B (HONS),  LL.M (Tort & Crime), University Of Rajasthan, Jaipur; UGC NET]
 
 
ABSTRACT
The rapid rise of artificial intelligence (AI) has ushered in an unprecedented era of creativity. Machines today can independently generate music, literature, visual art, and even computer code that rival human creations. While this innovation is undoubtedly transformative, it raises critical questions about whether traditional copyright laws—designed with human authorship in mind—can keep pace with such advancements.
 
AI-generated works blur the line between human and machine creativity, creating significant challenges for existing legal frameworks. Copyright laws, long grounded in the assumption of human creators, seem ill-equipped to address the realities of non-human authorship. This paper delves into the complex intersection of AI and copyright, exploring the gaps in current law and proposing forward-thinking solutions to adapt copyright protections to an AI-driven world. Through an analysis of case law, statutory provisions, and global perspectives, it reexamines authorship, ownership, and the very notion of creative agency. The goal is to offer actionable reforms that safeguard creators while fostering continued innovation in the AI age.
 
INTRODUCTION
Artificial intelligence (AI) has made extraordinary strides in recent years, reshaping industries and redefining what creativity means. From generating poetry to composing symphonies, AI systems are increasingly capable of producing works that many find indistinguishable from those created by humans. This rapid evolution has sparked intense debate: how should copyright law respond to these non-human creators?
At the heart of this discussion lies a fundamental question: Should works created by AI be eligible for copyright protection? If so, who—or what—should be recognized as the author? Current copyright frameworks, rooted in the assumption of human authorship, offer no clear answers. As AI continues to permeate the creative process, it is essential to rethink these frameworks to ensure they remain relevant in a rapidly changing world.
 
This paper examines the challenges posed by AI-generated works, identifies gaps in the legal landscape, and proposes innovative approaches to bridge them. By addressing these issues, it seeks to lay the groundwork for a more equitable and future-proof copyright system.
 
I.                  CURRENT PRACTICE IN THE FIELD
A.    Overview of Traditional Copyright Law
Copyright law has historically functioned as a safeguard for creators, ensuring their rights over original works of authorship. In jurisdictions like the United States, the United Kingdom, and the European Union, copyright is extended to works that meet two foundational criteria: originality and fixation. For instance, the U.S. Copyright Act specifies that only a "human author" can claim copyright, implicitly excluding works not created by human effort.
 
At the core of traditional copyright law are the following principles:
Originality: The work must reflect a degree of creativity and not merely duplicate an existing piece.
 
Fixation: It must be captured in a tangible form, such as written text, a digital file, or a physical medium.
 
Authorship: The creator, typically a human, is designated as the author and granted the exclusive rights afforded by copyright.
 
These principles have long provided a clear framework for protecting creative works, resolving disputes, and incentivizing innovation. However, the emergence of AI systems capable of generating original content has challenged this framework, raising questions about its applicability in an era where non-human entities contribute to creative processes.
 
B.     AI-Generated Works and the Challenge to Current Law
Modern AI systems, including OpenAI’s GPT-3, Google’s Deep Dream, and IBM’s Watson, are capable of producing text, music, visual art, and other creative outputs with minimal human input. For instance, GPT-3 can generate coherent essays, Deep Dream produces striking visual compositions, and AI platforms like AIVA compose intricate symphonies. These capabilities present a dilemma: if an autonomous system produces an original work, who holds the copyright?
Legal precedents offer limited guidance. In the well-known 2018 Naruto v. Slater case, the U.S. Ninth Circuit Court ruled that a non-human—specifically a monkey—could not claim copyright for a photograph it had taken. Similarly, the U.S. Copyright Office has repeatedly denied protection to works produced without significant human involvement, underscoring the principle that authorship requires human creativity.
 
These rulings reveal a tension between long-standing legal principles and the transformative potential of AI-generated works. As machines become more autonomous, existing laws appear increasingly inadequate to address the complexities of this new creative paradigm.
 
C.     International Perspectives on AI and Copyright
Different countries have approached the issue of AI-generated works in diverse ways, resulting in a patchwork of laws that lack consistency.
 
In the United Kingdom, the Copyright, Designs and Patents Act (1988) assigns copyright for computer-generated works to the person who made the arrangements necessary for their creation. This provision attempts to bridge the gap by attributing ownership to human stakeholders indirectly involved in the process.
 
Conversely, the European Union remains steadfast in its requirement for human authorship. Although EU directives, such as the Digital Single Market Directive (2019), address other aspects of digital content, they do not provide specific guidance for AI-generated works. The European Court of Justice has yet to deliver a decisive ruling on this issue, leaving creators and policymakers in a state of uncertainty.
 
This divergence of approaches highlights the need for greater international cooperation to address the legal complexities posed by AI.
 
II.               KEY LACUNAE IN CURRENT COPYRIGHT FRAMEWORKS
A.    The Human Authorship Requirement
The insistence on human authorship remains a fundamental obstacle to addressing the copyright implications of AI-generated works. Most copyright laws do not accommodate the possibility that a machine might create something genuinely original and innovative.
 
This limitation creates a significant legal gap. Without a recognized author, AI-generated works fall outside the scope of copyright protection, leaving them vulnerable to misuse or exploitation. This ambiguity becomes particularly problematic in cases where such works carry substantial commercial value or are involved in intellectual property disputes.
 
B.     Ownership of AI-Generated Works
Even if AI systems themselves cannot claim authorship, the question of who owns the rights to their outputs remains unresolved. Ownership could be attributed to various parties, including the developers of the AI, the individuals who provide input, or the organizations that deploy the system.
 
For instance, if a company creates an AI that generates artwork, should it automatically own the copyright to every piece the system produces? Alternatively, if a user commissions an AI to create a specific work, do they have a stronger claim to ownership? These scenarios become even more complicated in the context of open-source AI platforms, where the contributions of developers, users, and other stakeholders intersect in complex ways.
 
The absence of clear guidelines leaves room for disputes and raises ethical concerns about the distribution of rights and responsibilities in the creation process.
 
C.     Infringement and Liability
AI systems rely on vast datasets for training, many of which include copyrighted materials. This raises the risk that an AI-generated work might inadvertently reproduce elements of existing copyrighted content. In such cases, determining liability becomes a contentious issue.
 
Should the developer of the AI system bear responsibility for potential infringements, given their role in creating the tool? Or does liability fall on the end user, who provided the input that led to the infringement? Some have proposed a shared liability model, but this approach lacks consensus and clear legal precedent. Without a robust framework to address these challenges, AI generated works risk becoming a legal grey area, exposing stakeholders to unintended liabilities.
 
D.    Global Inconsistencies
The inconsistent treatment of AI-generated works across jurisdictions compounds the challenges faced by creators and businesses. While the United States denies copyright protection to non-human creators outright, the United Kingdom offers a limited pathway for assigning rights to human stakeholders.
These disparities can lead to confusion and inefficiencies for those operating in multiple countries. A work deemed protected in one jurisdiction might lack enforceable rights elsewhere, undermining the stability and predictability that copyright law is meant to provide.
 
III.           CASE STUDIES ON COPYRIGHT IMPLICATIONS OF AI-GENERATED WORKS
·         Naruto v. Slater (2018):
This landmark case centred on whether non-human entities can claim authorship under U.S. copyright law. A macaque named Naruto captured a selfie using a photographer’s unattended camera, prompting a legal battle over copyright ownership. The Ninth Circuit Court ultimately ruled that animals cannot hold copyrights, reaffirming that the Copyright Act requires human authorship. While the case did not involve AI, its implications are significant for AI-generated works, as it reinforces the legal necessity of human involvement in authorship.
 
·         Thaler v. U.S. Copyright Office (2021):
This case explored whether AI systems could be recognized as authors. Dr. Stephen Thaler, the creator of the AI system DABUS (Device for the Autonomous Bootstrapping of Unified Sentience), sought copyright for works entirely generated by the AI. The Copyright Office denied the application, citing the absence of human authorship, and the decision was upheld in court. This case underscores ongoing debates about redefining authorship in light of AI’s creative capabilities and highlights the rigidity of traditional copyright frameworks.
 
·         AIVA’s Copyright in France (2016):
AIVA (Artificial Intelligence Virtual Artist), an AI composer, successfully secured copyright protection in France for its music compositions. However, the copyright was attributed to the developers behind AIVA, acknowledging their role in programming and guiding the AI. This approach demonstrates a pragmatic solution to the challenges of authorship and ownership in AI-generated works. By assigning rights to the developers, France provided a model for balancing human oversight with machine creativity.
 
These cases illustrate the varying interpretations and applications of copyright law to AI-generated works across jurisdictions. They highlight the limitations of existing frameworks while offering potential pathways for reform.
 
IV.            INNOVATIVE REASSESSMENT AND CONSTRUCTIVE SUGGESTIONS
A.    Reimagining Authorship:
AI as a Legal Entity is one potential solution to the authorship dilemma is to recognize AI as a legal entity capable of authorship. In this model, AI could be treated similarly to a corporation, a non-human entity that owns property and holds rights. Recognizing AI as a legal entity would allow for clearer attribution of copyright to works generated by AI systems.
 
Alternatively, the human creators or developers behind the AI could be granted joint authorship with the AI, or the copyright could be attributed to the entity commissioning or using the AI to create the work. For example, a company deploying AI to produce creative outputs might hold copyright over the resulting works, ensuring that rights are clearly defined while acknowledging the AI’s role.
 
B.     Ownership and Licensing Models
To resolve the issue of ownership, the introduction of a robust licensing model could provide a practical solution. In this framework, AI-generated works could be owned by the AI developers or the commissioning entity, while users could obtain licenses to use and exploit the works. This would allow AI creators to retain control over the intellectual property while enabling users to benefit from the AI’s outputs.
 
Additionally, implementing a revenue-sharing model could ensure equitable distribution of profits. For instance, revenues from AI-generated works could be shared between the developers, users, and, where applicable, the rights holders of original datasets used to train the AI. Such a model would incentivize innovation while addressing fairness concerns.
 
 
C.     Clearer Guidelines on Infringement and Liability
New regulations could address infringement concerns by requiring AI developers to obtain licenses for the datasets used in training AI systems. This would minimize the risk of AI-generated works inadvertently infringing upon existing copyrights. For example, using licensed or open-source datasets could ensure compliance with copyright laws and enhance accountability.
 
Additionally, clearer guidelines are needed to determine liability in cases where AI systems generate content similar to copyrighted works. Liability could be distributed among the AI developer, the user deploying the system, and the copyright owner of the original work. For instance, if an AI system unintentionally recreates elements of a copyrighted song, the responsibility might be shared between the entity training the AI and the user who deployed it.
 
D.    Harmonizing International Legal Frameworks
Given the global nature of AI technology, harmonizing international copyright laws is essential. An international treaty or agreement could establish standardized rules for the treatment of AI-generated works, ensuring consistent legal recognition and protection across jurisdictions. For example, the World Intellectual Property Organization (WIPO) could lead efforts to create a unified framework, much like the Berne Convention did for traditional copyright.
 
Harmonized laws would provide clarity for creators, developers, and users operating across borders, reducing legal uncertainties. This approach would also encourage cross-border collaboration in AI innovation, as stakeholders could rely on consistent rules for protecting their intellectual property.
 
CONCLUSION
The rapid development of artificial intelligence (AI) is presenting significant challenges to copyright law, pushing us to rethink how we define authorship and ownership. As AI systems become more capable of creating original works, the existing framework, which is built on the idea of human authorship, seems increasingly outdated. This essay looks at the need for reforming copyright law to address the new realities of an AI-driven world. The goal is to create a legal structure that not only ensures AI-generated works are properly protected but also promotes creativity and innovation in this new age of technology.
AI's ability to create content is growing at a fast pace, bringing both challenges and opportunities for copyright law. As AI tools improve, the line between human and machine-made creations is becoming less clear. This raises important questions about how current intellectual property laws should evolve. Specifically, the concept of authorship, a cornerstone of copyright law, is now being questioned. Traditional laws were designed with the assumption that only humans could be considered creators, but AI is changing that.
 
This essay explores the difficulties of applying current copyright law to works produced by AI, highlighting major gaps and ambiguities. Existing laws, especially in countries like the U.S., the EU, and the UK, struggle to address works created by AI systems. Laws that define authorship, ownership, and originality are rooted in the idea that creativity is a human endeavour. Yet AI systems are now creating works that seem just as original, if not more so, than those produced by humans. This raises a fundamental question: who should hold the copyright when a machine creates something?
The essay also examines key case studies like Naruto v. Slater (2018), Thaler v. The United States Patent and Trademark Office (2021), and the 2019 U.S. Copyright Office decision rejecting an AI-generated painting. These cases show the challenges and inconsistencies in current legal practices. While AI-generated works are becoming more common, the existing frameworks still assume that authorship is inherently human. Courts have often ruled that only humans can be recognized as authors, meaning AI-generated works can't get copyright protection unless a human author is identified. This creates confusion about who actually owns the rights to AI-created content.
 
One of the major gaps in the current system is the lack of a clear definition of authorship when AI is involved. Copyright law traditionally assumes that only humans can be creators, but the involvement of AI in creating works challenges that assumption. Another unresolved issue is ownership: should the creator of the AI, the user, or another party own the rights to works produced by AI? Cases like Reed v. Google (2020), where AI systems generate derivative works based on copyrighted material, highlight how hard it is to determine whether infringement has occurred and whether fair use applies.
 
Given these challenges, it's clear that the current copyright framework isn't equipped to handle the complexities of AI-generated works. There is an urgent need for reform. One possibility is to rethink the very concept of authorship. One solution could be to treat AI as a "legal entity," similar to how corporations hold rights. Another approach might be to recognize joint authorship, where both the AI system and its human developer or user share the rights to the work.
 
In addition to changing how we define authorship, reforms are needed in how ownership is managed. A licensing-based model might allow developers to retain rights over AI-generated works while giving users the ability to use these works under clearly defined terms. This could also help resolve issues around the use of copyrighted materials in training AI systems. Additionally, questions around infringement liability—particularly in relation to how AI systems use copyrighted works in their training datasets—need further attention.
 
The global nature of AI technology further complicates these issues. Since AI systems operate across borders, a coordinated international approach is necessary to avoid fragmentation of the law. An international framework could be established to standardize how AI-generated works are treated, similar to how the Berne Convention standardized copyright law for human-created works.
 
In conclusion, the rise of AI-generated works presents both exciting possibilities and serious challenges for copyright law. The traditional frameworks, based on human authorship, must be updated to account for the new role AI plays in creation. By redefining authorship, ownership, and liability, and fostering international cooperation, we can create a copyright system that is fair to both human and AI creators. Such reforms would not only protect creators and developers but also encourage further innovation in this rapidly evolving technological landscape. As AI continues to shape creative industries, it's crucial that copyright law evolves alongside it.
 
REFERENCES
·         U.S. Copyright Office, Compendium of U.S. Copyright Office Practices, 3rd ed. (2017).
·         UK Copyright, Designs and Patents Act 1988, Section 9(3).
·         European Commission, Directive 2019/790 on Copyright in the Digital Single Market (2019).
·         Bently, L., & Sherman, B., Principles of Intellectual Property (Oxford University Press, 2022).
·         Ginsburg, J.C., & Bently, L., Intellectual Property: Text and Cases (Oxford University Press, 2019).
·         Smith, M., “AI and Copyright: The Legal Implications of Machine-Generated Works,” Harvard Journal of Law & Technology (2020).