AI-Generated Content and Copyright Infringement: A Legal Analysis of OpenAI’s Practices and Global IPR Challenges

Paper Code: AIJACLAV12RP2025

Category: Research Paper

Date of Publication: May 19, 2025

Citation: Ms. Abinaya Meyyammai Meyyappan & Ms. Kamalajyothini G, “AI-Generated Content and Copyright Infringement: A Legal Analysis of OpenAI’s Practices and Global IPR Challenges", 5, AIJACLA, 126, 126-133 (2025), <https://www.aequivic.in/post/ai-generated-content-and-copyright-infringement-a-legal-analysis-of-openai-s-practices-and-global-i>

Author Details: Ms. Abinaya Meyyammai Meyyappan, 3rd Year Law Student, Vellore Institute of Technology (Chennai Campus) &

Ms. Kamalajyothini G, 3rd Year Law Student, Vellore Institute of Technology (Chennai Campus)

Download PDF

Abstract

Generative AI technologies, such as OpenAI’s ChatGPT and DALL·E, are experiencing rapid growth. They have become the force which is disrupting the whole system of traditional intellectual property (IP) and have raised serious questions about authorship, ownership and the violation of rights. This paper is exploring the main challenges and dangers of AI-generated content in the legal and ethical spheres of copyright, trademark, and patent laws, and it explicitly points out the irrevocable weakness of the currently existing IP regimes tailored for human creativity. Focusing on the world at large, the research aims to shed light on the main questions that are: is it legal to use copyrighted works for the training of an AI system, can the results of the AI system violate the rights of protected works, and at what point the distinction between inspiration and replication becomes less and less clear with time. The cases of OpenAI and Perplexity AI, the lawsuits against these, perfectly serve as examples of the problems of adapting the copyright doctrines like fair use and substantial similarity which were originally meant for human-made works only to the AI-generated ones. Furthermore, there is a discussion of the issues that AI’s impact has on creative industries, in particular the paper proposes solutions to such problems, for example, the models of collective licensing and the obligation of transparency for the AI developers. This study, through the utilisation of a comparative legal study, has positioned the legal regimes of various countries such as the U.S., EU, UK, Indonesia, China, and India to highlight the weak points in the IP law and the need for international law adjustment. This paper primarily stresses the importance of the broadening of the copyright concepts so that they can be consistent with the AI-induced renaissance of creativity and economic growth while ensuring the usefulness of the working profiles of the parliamentarians and the judicial officers who are leading this field.

KEYWORDS: Generative Artificial Intelligence, Copyright Infringement, Collective Licensing, Fair Use Doctrine, OpenAI Source, Interoperability Standards

RESEARCH QUESTION

“How can we redesign intellectual property laws to address the challenges of AI-generated content while encouraging innovation, safeguarding human creativity, and ensuring ethical accountability?"

INTRODUCTION

"Artificial Intelligence (AI) is not really evil or good; it's simply a tool. What counts most is the way we decide to use it". These are the words, frequently repeated by technology leaders, that sum up the ambivalence of artificial intelligence. On the one hand, AI technologies such as OpenAI's ChatGPT and DALL·E have become icons of human achievement, producing art, writing essays, and even transforming industries. On the other hand, they force us to question our values. Are we, in trying to innovate, sacrificing the fundamental nature of human creativity?

Consider Sophia, the robot that was given citizenship. While millions of stateless people struggle for rights and recognition, what does it reflect about our values that we give such status to a machine? How did we arrive at this point? Is it because AI has filled in gaps where humans failed, or have we let ourselves rely too much on the convenience it offers?

AI has certainly changed lives—voice assistants such as Alexa and Siri have provided individuals with disabilities with a new level of independence, while applications such as ChatGPT assist students, scholars, and artists in finding answers quicker than ever before. But convenience comes at a price. As the distinction between inspiration and duplication becomes increasingly vague, are we cultivating creativity or encouraging copying? In a universe where AI can copy the Mona Lisa or compose poetry that rivals Keats, does it really matter who—or what—made it?

The creative industries, once strongholds of human ingenuity, are now in the hands of algorithms. Generative AI, learned on massive sets of human-created content, makes us squirm with questions about ownership and authorship. If an AI takes copyrighted material and creates something "new" out of it, who owns it? Can AI itself be credited as an author, or does the programmer behind the algorithm bear the responsibility? And if so, where does this leave the human creators whose work served as the foundation for the AI’s training?

These problems are not limited to copyright wars or legal principles. They are aimed at the very core of ethics and justice. Marginalized artists, whose work remains unseen, are disproportionately used as training data for AI. Oughtn't they be paid? Shouldn't developers be responsible for their systems? How do we maintain transparency and fairness in a field dominated by technology giants?

As we journey through this new bold world, the decisions we make today will determine the future of creativity. Will AI be used as a tool to enhance human potential, or will it displace it? The solutions lie in redefining intellectual property laws and promoting ethical responsibility. Because, after all, AI might be the future, but humanity needs to stay at its core.

1. TRAINING DATA: POOL OF KNOWLEDGE

                     I.            Legal Complications of Copyrighted Information in AI Training

Generative AI software, including OpenAI's GPT and Perplexity AI, work by processing massive datasets, which frequently include copyrighted content. This creates serious legal issues around fair use and intellectual property law violations. The issue is compounded by the unavailability of information from AI companies, which tend to keep information about their data sources secret, invoking trade secrets and competitive interests. It allows it to defy enforcement of copyrights and proper payments for original work creators[1].

                   II.             Assessing AI-Produced Content In Light of Transformative Use

The Andy Warhol Foundation for the Visual Arts Inc v. Goldsmith[2] court case established the essential precedent when analyzing transformative use. The Supreme Court held Warhol's taking of Goldsmith's photograph didn't qualify for fair use in light of little transformative change. This decision directly relates to AI-generated content, as most of it is based on copyrighted materials without going through substantial modification. In contrast to conventional creative transformation, AI-generated products frequently blur the distinction between derivative works and transformative use, thus complicating copyright enforcement.

                 III.            Fair Use vs. Copyright Infringement: A Precedent from the Google Books Case

The Google Books Project[3] is an important reference point for training data with AI. Courts decided that Google's mass digitization for search and indexing was fair use since it was done for a greater public interest. AI content, however, surpasses indexing in that it generates new, sometimes derivative work based on copyrighted material. Current legal contention hinges on whether AI training, such as that by Google Books, benefits the public at large or uses original work unfairly. Policymakers and the judiciary have to walk a fine line between promoting innovation and protecting intellectual property rights.

                 IV.            Ethical and Regulatory Issues Surrounding AI Usage of Data

AI systems are fed learning data that have bias, then that bias is translated into the media, advertising, and decision-making with discrimination. Regulatory frameworks such as the EU's General Data Protection Regulation (GDPR)[4] and the California Consumer Privacy Act (CCPA)[5] are really good at controlling the use of personal data, however, they are completely at a loss dealing with the peculiar problems brought by AI. It is the copyright owners who request the disclosure of AI training data sets in one of the fields, whereas the AI companies have been against it due to their right of trade secrecy. The operated laws of the EU and the U.S. will work on implementing the desired partial release of information so to balance the business requests and the transparency.

2. WHO RUNS THE WORLD? AI!

                     I.            The AI-driven Digital Divide and Copyright Concerns

AI-powered content creation disproportionately benefits large technology companies, further exacerbating the gap between resource-rich organizations and individual creators, especially those from marginalized communities. Limited access to sophisticated AI tools hinders chances for smaller creators, widening economic and creative divides[6]. At the same time, numerous artists and writers contend that AI-generated content infringes on their intellectual property without fair compensation. Existing royalty models are unable to distribute revenue fairly, with questions regarding whether creators do get to reap benefits from the digital economy.

                   II.             Authorship, Ownership, and Responsibility for AI-Generated Works

Ascertaining who owns and is responsible for AI-generated material continues to be an intricate legal conundrum. Existing copyright law, most specifically the "Work Made for Hire" doctrine, presumes human authorship with no established standards for works generated by machines. The Thaler v. Perlmutter[7] case reaffirmed that human authorship is a requirement for copyright protection, and thus purely AI-created works are not eligible. With the increasing role of AI in content creation, courts are still trying to determine the level of human contribution necessary for copyright eligibility and liability concerns, as in cases such as Thomson Reuters v. Ross Intelligence[8].

                 III.            Licensing and Copyright Enforcement in AI-Created Content

To handle fears of just compensation, economists suggest collective licensing models—like in the music sector—to pay original authors for AI-created works. But international cooperation would be required to make such an undertaking practicable. Copyright enforcement on AI-generated works is still problematic since AI can generate works that are almost identical to copyrighted works. Techniques like watermarking and content tracking are being researched in order to identify likely infringements. Courts continue to evolve liability models, establishing the role of AI developers and users in infringement cases.

                 IV.            Patentability of AI-Generated Innovations

The controversy surrounding AI inventorship has resulted in landmark cases across the globe. The DABUS AI cases, such as Thaler v. Vidal[9], reaffirmed that existing patent legislation identifies natural persons as inventors and not AI, excluding it from patent entitlement. Whereas UK, U.S., and Australian courts generally have abided by this line, debates are ongoing concerning the extent to which intellectual property regulations need to evolve to provide protections for innovations sparked by artificial intelligence. It is in a world context, courtesy of the World Intellectual Property Organization (WIPO)[10], that attention is now devoted to changing legislations with unique challenges regarding patenting resulting from AI applications.

3. INTERSECTIONAL ARBITRATION IN DEFENCE OF AI

                     I.            Transparency and Accountability: AI vs. Programmers

The intricacies of arbitrating AI-generated content or system disputes stem not only from the absence of well-defined legal frameworks but from how different stakeholders interact, i.e., developers, users, and affected third parties while juggling ethics and equity. One of the critical questions in AI disputes is whether to determine where the onus lies—on the AI system itself or its developers/programmers. Researchers such as Florian Möslein and Sebastian Zumbansen contend that AI, being a “black box,”[11] tends to be opaque, and it is hard to determine whether accountability rests with the algorithm or the choices of its developers. This problem is exacerbated when AI systems injure individuals, such as discriminatory hiring algorithms or navigation errors based on GPS that cause accidents.

Within an Indian perspective, the Digital Personal Data Protection Act, 2023 (DPDPA)[12] emphasizes accountability for transparency and data accountability by data processors that may include AI systems. Still, the enactment is not express regarding accountability over AI autonomous decision-making. Similarly, the Information Technology Act, 2000 (IT Act)[13], though foundational in addressing cyber issues, lacks provisions for AI-driven liability. This gap highlights the need for arbitration mechanisms that can fairly resolve disputes and ensure accountability.

                   II.            Ownership and Patents: Should AI Be Recognized as an Inventor?

The issue of whether or not AI-generated content can be copyrighted or patented as the inventor is yet another debatable topic. Cases like Thaler v. Comptroller-General of Patents[14] in the UK, where the court declined to recognize an AI system (DABUS) as an inventor, is a good example of the present limitations of intellectual property legislations. Critics contend that barring AI as an inventor suppresses innovation, whereas its supporters maintain that attributing inventorship to a non-human muddles the foundations of human creativity.

India's patent laws under the Patent Act, 1970 follow international standards that demand an inventor to be human and thus exclude AI from inventorship[15]. Arbitration mechanisms may resolve conflicts in joint ownership models where AI and its developers have intellectual property rights. Such models may be similar to the hybrid authorship model suggested by Ahmed Elgammal and others, who advocate for partial recognition of human contributors and AI systems[16].

                 III.             Collective Licensing and Oversight Mechanisms

Collective licensing presents a compelling solution to contentions regarding the use of copyrighted material by AI in its training data. AI developers would under this system remit licensing payments to use copyrighted material, giving human creators financial compensation. The European Union's Copyright Directive 2019/790 offers an initial point in that it takes up data-mining exceptions to AI systems but is limited to particular sectors only[17]. Arbitration would have a central role in enforcing collective licensing structures, resolving revenue-sharing conflicts, and promoting fair access to resources.

Yet, the absence of a regulatory body in India poses hindrances to effective regulation. The lack of arbitration and policy coordination mechanisms for settling disputes—from liability cases in autonomous traffic estimates to copyright infringement on AI-composed music—leaves actors without proper remedies. Researchers such as Daniel Gervais call for the institution of specialized AI arbitration panels to settle such disputes effectively[18].

                 IV.            Intersectionality in AI Disputes: Addressing the Marginalized

AI systems tend to be based on large datasets, which cover works protected by copyright that have been created by underrepresented creators and artists. Such works are generally employed without permission or proper remuneration, exacerbating economic and social disparities[19]. Arbitration should thus play the role of a mechanism to correct such injustices through equitable representation and compensation for all parties engaged in the creation and training of AI systems. Copyright legislation provides creators with exclusive rights to their work, but AI technologies tend to scrape data without permission, disproportionately subjecting marginalized creators. The failure to be open about how datasets are created allows these creators not to know they are misusing their work. Instances such as Getty Images v. Stability[20] AI call for fair remedies. Intersectional arbitration boards must incorporate representatives of these communities to promote fair compensation and representation.

Artificial intelligence systems can reinforce damaging stereotypes, usually perpetuating racial and bodily prejudices deeply rooted within societal institutions[21]. Most often, content generated by AI speaks to superficial characteristics, like color or body appearance, which have been the most overt signs of race in the past. For example, image recognition technology has been demonstrated to systematically misclassify individuals with darker skin, commonly labeling them with negative or subordinated traits on the basis of ingrained prejudice. These biases aren't only relevant to how AI systems "look" at the world—they also determine how AI engages with it, perpetuating conventional concepts of race and physicality as measures of value or ability. Additionally, in imaginative AI uses such as AI art, the same racialized tendencies tend to be exaggerated. AI images of certain ethnic groups tend to rely heavily on stereotypical indicators, e.g., darker skin, proportions, facial features, or specific attire, which will minimize rich cultural identity to overgeneralized and tokenistic depictions. By repeating and reinforcing such stereotypes, AI threatens to naturalize inbuilt racism, where prejudice towards particular groups becomes a natural result of machine logic instead of being an outcome of human programming. AI systems, if badly trained or not regulated, can enforce historical racial disadvantages under the pretence of neutrality and objectivity. As researchers such as Kate Crawford have identified, “AI systems are not neutral; they reflect the biases inherent in the data they are trained on and the people who create them”[22].

4. LOOKING AHEAD: AI, COPYRIGHT AND ETHICAL RESPONSIBILITY

                     I.            Finding the Balance: Innovation and Intellectual Property Protection

The paradox of AI copyright law is weighing technological progress with equitable protection to human creators. AI can drive revolution in innovation and creativity while endangering long-standing copyright infrastructures. Copyright frameworks need adaptation to accommodate technological advancements driven by AI while guarding equitable recognition and compensation for human originators. Scholars suggest implementing "AI-exclusive copyright carve-outs", which could delineate safe uses of copyrighted material in the training of AIs without crushing technological development[23].

                   II.            Corporate Responsibility and Transparency in AI Development

Perhaps the most urgent question is whether AI businesses must be legally liable for copyright infringement by their models. Proponents of liability contend that the businesses benefit from copyrighted material without paying creators, so licensing and dataset openness are critical to conformity. Others warn against excessive regulation, as it might impede AI advancement and economic growth. Balancing corporate responsibility with innovation promotion is still a core policy dilemma.

                 III.            Policy Reforms for a Fairer AI-Copyright Landscape

To facilitate ethical and legal AI development, policymakers need to consider thorough legal reforms. The major suggestions are:

●        Mandatory transparency: Companies making AI must reveal datasets utilized for training.

●        Fair compensation models: A collective licensing model might give original creators royalties on AI-created works.

●        Global AI copyright standards: An international treaty may unify copyright protection across the globe, minimizing differences among jurisdictions.

These steps will try to bring copyright law into the modern age and promote both human creators and AI innovation[24].

                 IV.            The Future of AI and Copyright: Towards Ethical AI Development

As AI increasingly alters creative industries, there is an urgent need to determine moral responses to AI-produced content. Governments, technology companies, and artists will need to collaborate on establishing sustainable regimes for copyright. The destiny of AI copyright depends on the ability to let innovation flourish without undermining the rights of human creators, promoting an equitable and transparent digital environment.

CONCLUSION

AI-generated content has brought tremendous transformation in the classical methods of applying copyright law, leading to enormous challenges for legality as well as ethical and creativity issues. Getty Images v. In the Stability AI v. New York Times. The examples of OpenAI & Microsoft showcase the growing friction between AI creativity and human ingenuity, prompting questions about who is actually behind it. Is it right to blame AI systems for copyright violations, considering that they are trained on huge datasets that are often scraped without permission? Can AI be regarded as an inventor, as the DABUS case argues, or is it solely based on human work for authorship? The questions highlight the lack of current intellectual property models, which are insufficient in handling the details of AI. The global regulatory landscape is divided, with the EU's rigorous AI Act, the U.S.' more elastic fair use doctrine, and Asia-Pacific' rigid methods creating a fragmented system. An integrated global solution, in the form of an AI-based copyright treaty led by WIPO or UNESCO, which would bridge these gaps and provide uniform enforcement across borders, is necessary.

This argument is all about the requirement to find a balance between innovation and responsibility. Inequalities are further worsened by insufficient transparency of AI training data, paired with the digital divide that is disconnecting billions and making them susceptible to disruptions. Mandates like compulsory attribution frameworks and collective licensing models are promising routes to gain just compensation at the cost of making progress in the growth of AI. Can we employ the transformative powers of AI without desecrating human cultural and creative achievement? How might this be accomplished if possible? The laws of copyright are obligated to enshrine maximum transparency, balance, and a sense of rightness at the critical turn here. By adjusting intellectual property systems to address these concerns, we can create an innovation-led mechanism that uphold the originality of human imagination in the face of mounting influence and dominance by AI. What we choose today will influence the cultural and legal context of tomorrow.

[1] Andres Guadamuz, ‘Talkin' 'Bout AI Generation: Copyright and the Generative-AI Supply Chain’ (2023) arXiv https://arxiv.org accessed 15 February 2025.

[2] Andy Warhol Foundation for the Visual Arts Inc v Goldsmith [2023] USSC 6.

[3] Google Inc v American Authors Guild [2015] 804 F.3d 202 (2nd Cir).

[4] Regulation (EU) 2016/679 (General Data Protection Regulation).

[5] California Consumer Privacy Act 2018.

[6] International Telecommunication Union, ‘Measuring Digital Development: Facts and Figures’ (2023) https://itu.int accessed 15 February 2025.

[7] Thaler v Perlmutter [2023] US District Court, District of Columbia.

[8] Thomson Reuters v Ross Intelligence [2023] US District Court, District of Delaware.

[9] Thaler v Vidal [2022] 43 F.4th 1207 (Federal Circuit).

[10] World Intellectual Property (WIPO), ‘AI and Intellectual Property’ (2023) https://www.wipo.int/ai-ip accessed 18 February 2025.

[11] Florian Möslein and Peer Zumbansen, ‘The Black Box of AI: Transparency and Accountability in Algorithmic Decisions’ (2021) 12(1) European Journal of Risk Regulation 17.

[12] Digital Personal Data Protection Act 2023 (India).

[13] Information Technology Act 2000 (India).

[14] Thaler v Comptroller General of Patents Trade Marks and Designs [2020] EWHC 2412 (Pat); see also Journal of Intellectual Property Law & Practice (2020) 15(10) 753.

[15] The Patents Act 1970 (India), s 2(1)(y); see also N Saha and A Bhattacharya, 'Artificial Intelligence and Patent Law in India: A Legal Analysis' (2021) 26(1) Journal of Intellectual Property Rights 33.

[16] Ahmed Elgammal and others, ‘Can Artificial Intelligence Be an Author? A Hybrid Approach to Authorship in the Age of AI’ (2022) 73 Journal of Artificial Intelligence Research 1.

[17] Directive (EU) 2019/790 of the European Parliament and of the Council of 17 April 2019 on copyright and related rights in the Digital Single Market [2019] OJ L130/92.

[18] Daniel J Gervais, 'The Human Cause: Artificial Intelligence, Authors and the Copyright Ecosystem' (2021) 23(1) Vanderbilt Journal of Entertainment and Technology Law 1.

[19] K Johnson, ‘The Exploitation of Marginalized Creators: The Impact of AI on Copyright’ (2019) 2 Journal of Technology and Social Justice 48.

[20] Getty Images (US), Inc v Stability AI Ltd (2023) Case No 23-c Getty Images (US), Inc v Stability AI Ltd (2023) Case No 23-cv-00135 (D Del).v-00135 (D Del).

[21] Ruha Benjamin, Race After Technology: Abolitionist Tools for the New Jim Code (Polity Press 2019).

[22] Kate Crawford, Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence (Yale University Press 2021).

[23] Daniel Gervais, ‘AI Creativity and Copyright Balance’ (2022) Journal of the Copyright Society of the USA.

[24] William McGeveran, ‘Corporate Accountability in AI Content Creation’ (2023) Harvard Business Review.