AI's Impact On Copyright Law - Legal, Ethical, And Practical Considerations
Author: LawSphereX
Category: Intellectual Property
Published:
Updated:
AI'S IMPACT ON COPYRIGHT LAW: LEGAL, ETHICAL,
AND PRACTICAL CONSIDERATIONS
Yashvir Maurya, (Intellectual Property), Amity University, UP, Amity Law School Noida Amy Kannan, (Intellectual Property), Amity University, UP, Amity Law School Noida Teli Tekam, (Intellectual Property), Amity University, UP, Amity Law School Noida
ABSTRACT
Artificial Intelligence (AI) is transforming multiple sectors such as entertainment, healthcare, education, and technology, raising complex legal issues surrounding copyright and authorship. As AI becomes increasingly involved in producing creative works, there is no consensus on how to address AI-generated content within current copyright structures. Three key positions have emerged: (i) granting AI the status of an author under copyright law, (ii) categorizing AI-generated works as public domain content, and (iii) proposing a new legal framework specifically designed to address these works. The debate centres around whether non-human entities like AI should enjoy the same rights as human creators, or if its creations should remain open to public use. Traditionally, copyright law is meant to reward human creativity, but its application to AI-generated works is complicated.
Advocates for recognizing AI as an author argue that machine-generated creativity should be valued similarly to human creativity. However, critics caution that this could undermine the value of human authorship and stifle innovation. Additionally, AI-generated content could be flawed, biased, or even harmful, leading to legal concerns about liability, particularly when such content causes defamation or incites violence. The issue of the copyright duration also becomes problematic, as AI can produce an unlimited number of works without experiencing fatigue or aging, challenging the traditional humanbased approach. The Indian case Amar Nath Sehgal v. Union of India highlighted that moral rights are inherent to human authors and cannot apply to AI. Similarly, in Camlin Pvt. Ltd. v.
National Pencil Industries, the court ruled that a machine cannot be considered an author, reinforcing the notion that copyright protection is for natural persons, not AI. International treaties such as the TRIPs Agreement and Berne Convention do not address non-human authorship, but they allow countries to implement
stronger protections. Moreover, technologies like deep fakes present additional legal challenges, such as privacy violations, defamation, and unauthorized usage. A practical solution to these challenges involves licensing agreements that
give AI developers permission to use copyrighted content while ensuring creators are compensated. Additionally, some platforms are implementing terms of service to prevent the misuse of AI-generated works and protect intellectual property rights. Keywords: Artificial Intelligence, Copyright Law, AI-Generated Works, Authorship, Public Domain, Sui Generis, TRIPs Agreement, Berne Convention, Deep Fakes, Licensing, Intellectual Property Rights, Case Law. AI and its impact in Copyright Law: An Overview Artificial Intelligence (AI) has gained considerable significance in modern times, becoming essential in a wide range of technological applications.
AI has transformed our lives by entering into various sectors such as heath, transportation, aviation, space, education, entertainment industry (music, art, games, films) and others and a tendency has been developed in all countries to automate most of the activities and minimize human intervention to ensure efficiency and rule out errors1. The term "Artificial Intelligence" was introduced by John McCarthy in 1956. Currently, there is no legal definition for "artificial intelligence. " It can be described as "the ability of machines to perform tasks that would typically require human intelligence. " In 1990, Ray Kurzweil defined AI as "the science of making computers perform tasks that would require intelligence if done by humans"2.
There are three main viewpoints on AI-generated works and copyright: (i) AI should be recognized as an author under copyright law, (ii) AI-generated work should not have authorship and should fall into the public domain, and (iii) a sui generis law should be established to protect such works. Copyright law typically incentivizes human creativity, but if AI is treated as an author, it could imply that "human creativity" and "machine creativity" are equally valued, potentially diminishing human creativity in the long run. On the other hand, if AIgenerated works aren't protected, it would prioritize human creativity over machine-generated 1 V. K. Ahuja, Contemporary Developments in Intellectual Property Rights: A Prologue, in V. K. Ahuja & Archa Vashishtha, Intellectual Property Rights: Contemporary Development 3, 3-18 (Thomson Reuters 2020).
2 Author(s), Impact of Artificial Intelligence on Copyright Law: Challenges and Prospects, RESEARCHGATE, https://www. researchgate. net/publication/377334695_Impact_Of_Artificial_Intelligence_on_Copyright_Law_C hallenges_and_Prospects.
output. Considering AI as an author brings several complications. AI-generated works may be flawed or harmful, using biased or toxic language, leading to issues such as defamation or incitement to violence. Holding AI liable for these issues is difficult as it lacks personhood. In extreme cases, the work may be deleted or the AI software banned, but irreversible harm could already have occurred. Additionally, if AI's work resembles existing copyrighted content, it raises questions about how to handle potential infringement. Furthermore, treating AI as an author would prevent it from transferring ownership, as it lacks legal personhood, creating further legal challenges. The copyright laws in many countries grant moral rights to authors, even though this is not required under the TRIPs Agreement.
Typically, two moral rights are recognized: (i) the right of attribution, and (ii) the right of integrity. The right of attribution ensures that the author is credited and associated with their work, while the right of integrity allows authors to seek compensation if their work is altered or distorted in a way that harms their honour or reputation. In Amar Nath Sehgal v. Union of India3, The Delhi High Court stated that "in the material world, laws are designed to safeguard the right to fair compensation. However, life extends beyond the material and t is also temporal. Many believe in the soul, and the moral rights of the author represent the soul of their work. The author has the right to preserve, protect, and nurture their creations through these moral rights. " These moral rights are tied to the emotions and feelings of the human author and are not applicable to AI.
A key issue regarding AI-generated works is the duration of copyright protection. Since AI
does not age or experience fatigue like humans, it can produce an unlimited number of works. Some argue that copyright for AI works is questionable, as human authors create a limited number of works within their lifetime, and their efforts are rewarded through copyright protection. In contrast, AI is immortal and can generate the same output repeatedly if given the same inputs. Critics of granting copyright to AI-generated works claim that such creations lack uniqueness and creativity, making copyright protection less justifiable. It would be challenging for AI to negotiate royalties or enforce rights under copyright law, as AI cannot engage in such activities. Making AI the author of a work could create more complications than solutions. Another perspective is that AI-generated works should not have authorship and should fall into the public domain.
This is justified because AI incurs no costs in creating works, and since AI
3 2005 (30) PTC 253 (Del.).
can produce limitless iterations without additional resources, it doesn't need economic or moral incentives to create. However, without protection, companies investing heavily in AI could be undermined by others exploiting these works for free, making some protection necessary to sustain further investment in AI development. The Indian Copyright Act does not define “computer-generated work” like CDPA. It however defines “author” in relation to “any literary, dramatic, musical or artistic work which is computer-generated” as “the person who causes the work to be created”4. In Camlin Pvt. Ltd. v. National Pencil Industries5, The Delhi High Court provided clarification on the term "author," stating that a "mechanically reproduced printed carton" cannot be copyrighted because it is impossible to identify the author of such a work.
The court explained that copyright protection is only granted to natural persons who create original works. As a result, the plaintiff cannot claim copyright over a carton produced through mechanical reproduction, as it is not the creation of an author. The court concluded that a machine cannot be recognized as the author of an artistic work, nor can it possess copyright over it. The idea of considering AI and human authors as joint authors is flawed, as humans do not control all AI operations. According to the Indian Copyright Act, joint authorship requires collaboration, which doesn’t apply to AI. Similarly, making AI programmers or users joint authors of AI-generated works is also problematic. For example, the Indian Copyright Act,
1957 defines “work of joint authorship” to mean “a work produced by the collaboration of two
or more authors in which the contribution of one author is not distinct from the contribution of the other author or authors”.6 At the international level, the Berne Convention, 1886 did not contemplate “non-human authorship”7. The same position applies to the TRIPs Agreement, as it incorporates provisions from the Berne Convention. The WIPO Copyright Treaty and WIPO Performances and Phonograms Treaty also support this view. However, international copyright law does not rule out the possibility of non-human authorship in national laws, allowing countries to offer greater protection than the minimum standards outlined in these treaties. In addition to authorship concerns, WIPO has identified a new issue regarding "deep fakes." Deep fakes involve creating
4 The Copyright Act, 1957, § 2(d)(vi) (India)
5 AIR 1986 Delhi 444.
6The Copyright Act, 1957, s.2(z).
7 Sam Ricketson, People or Machines: The Berne Convention and the Changing Concept of Authorship, 16
COLUM. VLA J. L. & ARTS 1 (1991).
simulated likenesses of individuals, including their voice and appearance, often using AI. This raises not only copyright issues but also concerns about privacy, defamation, and unauthorized use. Deep fake content featuring celebrities, performers, or leaders could become widely popular and generate significant revenue, potentially continuing even after their death, which complicates the legal and ethical landscape. Practical considerations: Due to the legal complexities of fair use and the concerns of copyright owners, licensing copyrighted data for AI training has been proposed as a possible solution. Licensing agreements could grant AI developers explicit authorization to use copyrighted content for training, while also ensuring that copyright owners are compensated for the use of their intellectual property. Content creators have a few practical options beyond copyright law.
One option is for artists to argue that AI-generated songs resemble deep fakes and infringe on their right to publicity. In its lawsuit against Stability AI, Getty Images has alleged the AI-generated images infringe upon the company’s trademarks. Getty Images protects its digital assets by including a company trademark watermark on each image8. AI-generated works that incorporate trademarks alongside low-quality content can damage and weaken the brand’s reputation. Additionally, publishers can take advantage of restrictions in usage agreements to prevent AI companies from scraping their content for training purposes. Websites such as Amazon, the New York Times and Shutterstock now include provisions against content scraping and data mining in their terms of service.
9 Similarly, platforms such as Getty Images have taken proactive measures to prevent the uploading of AI-generated works through their terms and conditions. The risk of AI algorithm users inadvertently triggering various legal liabilities means that employees and contractors need to collaborate across functions to define acceptable use cases, specify which company data can be used for training algorithms, and determine which platforms are allowed. To steer clear of legal issues, it's also recommended to obtain permission or licenses from copyright owners before utilizing protected content. The impact of AI on copyright law raises several practical issues. As AI systems generate content, 8 James Vincent, Getty Images Sues AI Art Generator Stable Diffusion In The US For Copyright Infringement, THE VERGE (Feb. 6, 2023), https://www. theverge.
com/2023/2/6/23587393/ai-art-copyright-lawsuit-gettyimages-stable-diffusion (accessed March 20,2025). 9 Amanda Lee, Websites Like Amazon, New York Times Block OpenAI’s Web Crawler From Obtaining Content, THE STRAITS TIMES (Aug. 25, 2023), https://www. straitstimes. com/tech/websites-like-amazon-new-york-timesblock-openai-s-web-crawler-from-gaining-content (accessed March 20, 2023).
questions emerge about authorship, ownership, and the use of copyrighted materials for training. Content creators and copyright holders must address potential violations, such as AIgenerated works infringing on rights to publicity or damaging trademarks. Licensing agreements can offer clarity, enabling AI developers to use copyrighted materials while compensating creators. Platforms like Getty Images also restrict AI-generated uploads. To avoid legal challenges, companies must establish clear guidelines on acceptable use, secure permissions, and seek licenses before utilizing protected works in AI training. Licensing Copyrighted Data for AI Training: Challenges and Emerging Solutions The rapid growth of Artificial Intelligence (AI) has raised significant challenges regarding its relationship with copyright law, especially in terms of AI training.
AI systems require access to large datasets, including copyrighted materials, to enhance their functionality. As AIgenerated content becomes more common, addressing the legal complexities associated with using copyrighted data for AI training has become increasingly important. Licensing agreements offer a potential solution to ensure creators are compensated for their work, while enabling AI developers to use these resources for training purposes. One of the main obstacles to licensing data for AI training is the vast amount of copyrighted content needed. AI models, especially those based on machine learning, depend on diverse datasets to improve their performance. This presents challenges in obtaining permission for using such extensive data, particularly when the content is owned by various copyright holders across different sectors.
Traditional licensing agreements may not scale well to accommodate the enormous volume of data required by AI systems. Moreover, issues arise regarding the nature of AI-generated works. AI systems can produce content that closely mimics or even copies the copyrighted material they were trained on. This could lead to potential copyright infringement concerns, as creators may not have consented to the use of their works for training AI or received compensation for the reproduction of their work in AI outputs. One possible solution is the development of standardized licensing agreements between AI developers and copyright owners. These agreements could grant AI developers permission to use copyrighted works for training while ensuring that copyright holders are appropriately compensated.
Licensing platforms or intermediaries could facilitate this process by creating contracts that streamline negotiations and make the licensing process more efficient. Some platforms and content creators are taking proactive steps to prevent unauthorized AI
training. For instance, Getty Images has taken legal action against Stability AI, alleging unauthorized use of its images for training purposes. Other companies, such as Amazon and The New York Times, have incorporated clauses in their terms of service to prevent AI companies from scraping their content. These protective measures help ensure intellectual property rights are respected. Additionally, some AI developers are exploring alternatives such as using public domain data or data specifically made available for AI purposes, which could reduce copyright infringement risks. As AI technology continues to evolve, the intersection of copyright law and AI training remains a key concern. Licensing agreements present a way forward to balance the interests of copyright holders and AI developers.
However, this solution must be tailored to the complexities of AI technology to ensure fair compensation for creators while fostering innovation in AI development.
PROPOSED SOLUTION AND CONCLUSION
The integration of Artificial Intelligence (AI) into various industries presents both substantial opportunities and significant challenges, particularly in relation to copyright law. As AI increasingly generates creative works and relies on copyrighted data for training, it raises complex questions about authorship, intellectual property rights, and the use of such data. To address these issues, several potential solutions can be implemented to strike a balance between encouraging technological innovation and safeguarding the rights of creators and copyright holders. One of the most pressing issues in the relationship between AI and copyright law is the need for AI systems to be trained using large datasets that often include copyrighted content.
To address this, licensing agreements can be developed that allow AI developers to use copyrighted works for training purposes while ensuring that the creators of those works are compensated. These agreements would grant explicit permission for the use of content, providing an efficient way to balance the interests of copyright holders and AI developers. However, given the scale of data required for AI training, it is essential that these agreements be standardized across industries. Licensing platforms or intermediaries could facilitate the negotiation and execution of these contracts, helping to streamline the process and ensure fair compensation for creators.
The current copyright framework does not address the unique challenges posed by AIgenerated works. To resolve this, AI-specific copyright laws could be established, clarifying the status of AI in the creative process. Such laws would need to address whether AI can be recognized as an author or whether human creators should retain ownership of works generated by AI. This framework would also need to outline how such works are protected and managed. Additionally, provisions could be made for situations where the AI-generated works are based on copyrighted content, ensuring that appropriate permissions are granted and compensation provided. These specialized laws would help provide clarity and consistency in the treatment of AI-generated content within the copyright system.
Another potential solution is to expand the concept of fair use, particularly for non-commercial, educational, or research purposes. This could allow AI developers to use copyrighted materials for training without the need for licensing, provided certain conditions are met. Additionally, encouraging the use of public domain content or datasets made specifically for AI purposes would reduce reliance on copyrighted works. This could help mitigate legal challenges related to the use of protected materials in AI training and ensure that developers have access to the necessary resources for innovation. To protect against unauthorized use of copyrighted works in AI training, content creators can implement proactive measures, such as including clauses in terms of service agreements that restrict AI developers from scraping their content.
For example, platforms like Getty Images and The New York Times have included specific provisions in their terms of use to prevent the unauthorized use of their works. Furthermore, employing technologies such as watermarking or digital rights management (DRM) could help track how content is being used and protect against unauthorized reproduction or use in AI systems. Alongside legal solutions, it is essential to establish ethical guidelines to address concerns such as bias, privacy violations, and misinformation associated with AI-generated content. These guidelines should ensure that AI systems are developed responsibly and that the works they produce are not harmful or misleading.
Accountability mechanisms should be put in place to hold AI developers responsible for the actions of their systems, particularly in cases where AI-generated content leads to harm, such as defamation or spreading false information. Regulatory bodies could be created to oversee AI development and ensure that ethical standards are upheld.
CONCLUSION
The intersection of AI and copyright law presents a range of legal and ethical challenges that
need to be addressed to foster innovation while protecting creators’ rights. Licensing agreements for AI training data, the introduction of AI-specific copyright laws, and expanding fair use provisions are potential solutions to the complexities of AI-generated works. At the same time, preventing unauthorized use of copyrighted content and implementing ethical guidelines will be crucial in ensuring that AI’s impact remains positive and constructive. A multifaceted approach, involving legal reforms, industry collaboration, and ethical oversight, will be necessary to adapt to the rapid development of AI technology while ensuring that intellectual property rights are respected and the interests of content creators are safeguarded. AI's growing role in creative sectors raises complex issues in copyright law, especially regarding authorship and ownership of AI-generated works.
Current frameworks struggle to address whether AI or its creators should hold rights over such works. While placing AIgenerated content in the public domain may discourage further investment in AI, offering full protection to these works could overshadow human creativity. The World Intellectual Property Organization (WIPO) is working on solutions, with a sui generis system or specialized copyright provisions for AI being potential options. A balanced approach, prioritizing human creativity while recognizing AI's contribution, is essential for future regulations.