I. INTRODUCTION
The evolving narrative of artificial intelligence (AI) brings to mind an imaginative concept portrayed in a popular 90s TV show, i.e. Shaka Laka Boom Boom, where a magical pencil transformed drawings into reality. While such fictional ideas entertained the masses, advancements in AI today have materialised many previously unthinkable technologies; automated systems, national identity databases (Aadhar in India, b-Cadastros in Brazil), and biometric tools are not magical or predictions but integral to modern life.
However, the implications of these advancements, particularly on intellectual property rights (IPR), remain a subject of extensive debate. Pioneers of artificial intelligence remain uncertain about its long-term effects on ownership, patentability, and legal frameworks, leaving much to speculation that when AI converges with human creativity and artistic expression, the understanding of these traits we have traditionally considered to be uniquely humane is radically transformed. As a result, the development of AI since the 1970s has faced hurdles like the “AI winters” caused by financial and technological limitations. A flexible framework may well be designed that goes around the issues that AI has raised, in keeping with a futuristic and welfare-oriented understanding of AI and humanity, without overstepping each other’s boundaries.
II. WHO OWNS IT?
The rise of generative AI raises profound questions about ownership and originality in creative works. These systems, capable of producing content indistinguishable from human creations, challenge traditional frameworks of intellectual property. Society thrives on innovation, with IPR playing a pivotal role in incentivizing creativity. IPR grants creators exclusive rights to their inventions, ensuring protection against unauthorized use while motivating further advancements. The field of art has been no stranger. These issues have significant implications for copyright law, which has historically tied authorship to human creators and the human “touch” (Baker Donelson).
At the heart of this debate lies the question: if AI replicates not just works but the creative process itself, who owns the output, and can it ever be considered truly original? Can a creation through a creation hold ownership rights? Is it Art-istic or Art-ificial?
Generative AI tools, including advanced language models and image-generation software, have revolutionised creative processes. These tools analyse existing datasets to generate new outputs, often with a precision that rivals human artistry. This technological capability disrupts the assumption that individuals inherently own their creative processes. Historically, intellectual property laws protected the outputs of individual intelligence. However, with AI imitating the intricate methods of creation, these protections may need to expand to include not just the creations themselves but the processes and styles that define human ingenuity.
The case of Naruto v. Slater, which saw a legal battle over copyright for a “selfie”, was one of the early cases that mulled over the extent of non-human entities’ – particularly a monkey’s- rights in legal ownership. However, this decision, based on the biological inability of animals to intend creativity, fails on many logical grounds when applied to AI.
III. THE CHALLENGE OF NOVELTY AND EVOLVING INTERPRETATIONS OF CREATIVITY
The privilege of having protection over creation through IPR is predicated upon the notion that a general novelty, non-obviousness and industrial application characterise it. The question of human imitation v. AI replication plays a crucial role. Existing styles, including personal talent, interpretation, and a unique perspective, inspire human creativity. In contrast, AI can precisely replicate creative styles, often producing results nearly indistinguishable from the original. This capacity for exact and scalable reproduction moves the dynamic from transformative inspiration to mechanical duplication, raising fundamental questions about originality. Therefore, can art generated by AI truly be novel?
The artwork “Edmond de Belamy”, a GAN creation that sold for a staggering $432,500, demonstrated the possibility of AI as a valid creator. The “digital” tulips of Anna Ridler, art created by Google AI, and The Machine Vision series are yet more examples of the use of AI in art. Cases like this have opened Pandora’s box when it comes to AI and art: (a) In the case of a generative art created by a particular AI model, will it be considered an infringement of IPR if another model(s) of AI possessing similar algorithms of functions and properties replicated the artwork? (b) Can the element of human contribution to generative art be considered a requirement for novelty that differentiates the algorithms in several types of AI? (c) Can it be possible to enumerate the extent of human ingenuity in devising the prompt about the process of creating art? (d) Is it possible for AI algorithms trained on pre-existing datasets to create original output? The present system of the IPR framework on the global level fails to address the complexities of such issues.
The case of a debate over copyright registration of an artwork created using an AI-based tool, RAGHAV, in India prompted the idea that AI is merely a tool, perhaps like the paintbrush the artist uses in her work. Though not strictly related to art, the Beijing Internet decision seems to stem from similar logical conclusions, where the voice, a personalised form of expression, was held to be patentable despite the use of AI.
Notably, the decisive factor was human benefit in further judgments like the Zarya of the Dawn case or the DALL-E Open-AI case. The reasoning consistently reiterates that a fine line is maintained between human and non-human entities and neglects the question of what artistic purposes an artwork created by AI serves.
Human creativity has often manifested in unconventional ways and has frequently collaborated with non-human entities. Artworks of Congo the chimpanzee, Andy Goldsworthy’s art that usually fuses creativity with the environment, and Patrick Tresset’s robotic drawing entities challenge the notions of process in art or how an artwork may be created. It is therefore necessary to consider the creations of AI in an artistic sense and what may be defined as personal in the age of the fourth industrial revolution.
IV. LEGAL AND ETHICAL AMBIGUITIES IN THE CURRENT INTELLECTUAL PROPERTY FRAMEWORK
Current intellectual property laws offer limited clarity on the ownership of AI-generated content. In some jurisdictions, works entirely generated by AI without human intervention fall into the public domain, as they lack a clear claimant. Where human input guides the AI, the degree of creative contribution necessary to establish ownership remains undefined. Moreover, pre-existing works for training AI systems continue to spark debate under fair use doctrines.
Issues are magnified due to the different interpretations of judicial systems. The United States IPR law continues to hold that humans should author the work to receive protection, as echoed in the recent case of Thaler v. Perlmutter. However, the UK treads a more flexible path wherein AI may use publicly available content for training under contractual agreements (UK IPR Law). In 2021, India recognised AI as a co-author in artistic creations but continues sceptical of property rights, as seen in the RAGHAV case.
Organisations like the European Union Intellectual Property Office (EUIPO) have done studies to establish that AI is not only a tool for enforcing IP laws but also an agent of violation. The EU has made some initial progress in developing a regulatory framework that extends to issues like developers’ transparency obligations, compliance policies, “high-risk” AI, text, and data mining. The EU AI Act entered into force this year and on a deeper observation we find, that it gives a rather significant Frankenstein complex with respect of AI.
AI assumes datasets which brings the issues of consent, transparency, compliance and privacy to the fore. Generative AI tools like ChatGPT and Gemini are known to rely on unauthorised data to answer user prompts, skewing the basis of originality. While the EU Act’s overarching scepticism does not justify these issues, the algorithmic bias, deepfake artworks, and transparency in AI workings do not exacerbate the systemic dilemmas, thanks to the “black box” nature of most AI models.
AI tools are often trained using copyrighted artworks of smaller or lesser-known artists without consent or attribution. Then these artworks are monetised to benefit large-scale and powerful companies (Istart law). The same goes for Indigenous artists, where cultural appropriation presents a critical problem. AI systems are also no less susceptible to promoting discrimination by often reflecting mainstream and dominant cultural narratives, ignoring the realities of marginalised communities.
V. RECOMMENDATIONS: NEED FOR AN INCLUSIVE REGULATORY FRAMEWORK
The rapid evolution of AI has called for the development of a global regulatory framework, as World Intellectual Property Organization (WIPO)’s recent issues paper especially highlights. WIPO has deliberated upon the role of data in AI training, requiring global measures for equitable sharing and diverse datasets due to cross-border data sharing. Addressing the North-South divide is critical, since high-income nations happen to dominate AI research as developing countries face systemic barriers. In order for technology policy development to occur transparently, accountability of the policy processes is needed to promote fairness, definition, and traceability of AI systems, like WIPO suggests for developing trust and mitigate risks.
WIPO’s approach sets out important issues, but there will be very significant challenges of implementation. The focus on data diversity is certainly warranted, as inclusive datasets are necessary to deal with systemic biases and increase applicability beyond the official data producing countries. However, the framework is in lacking of actionable mechanisms for nations with limited data resources, as these measures are overly reliant on the cooperation of high-income countries, which historically comes with conditions that risk perpetuating dependency rather than fostering empowerment. Similarly, addressing the North-South divide is a commendable focus, with technology transfer and capacity-building initiatives. However, the framing completely leaves out structural power disparities (like the disproportionate impacts on small creators, as well as the digital divide that continues to prevail across countries) which govern the distribution of resources, which undermine these types of initiatives. It will also be necessary to consider the future of artistic innovation and creativity in the light of AI. WIPO’s approach as it stands is more aspirational than transformative.
Additionally, WIPO’s role as a neutral convener is a vision for a platform that harmonizes AI regulations, geopolitical tensions and conflicting national priorities. Contradictory approaches reflected through the EU’s strict regulatory stance versus the US’s innovation-first approach, significantly slow down meaningful collaboration. WIPO must institutionalise collaboration through a tiered regulatory approach in the form of binding multilateral agreements that mandate equitable resource-sharing with enforceable mechanisms to ensure compliance.
VI. CONCLUSION
We need to constructively think about a respectful co-existence between machine-produced artwork and human creativity. The human-machine divide can blur and are not easy to define which complicates the law of ownership. Even so, we are reminded that creativity is not a defined point in time and is ever-evolving in terms of social and cultural standards. The introduction of machine-produced or AI artwork only adds space and options for artistic expression in general, and we must redesign our intellectual property regulation to reflect the newly constructed space. These adjustments have to not only include the structural issues but also fully support ethical ownership considerations with respect to AI generated artwork.
Authored By: Ms. Indu Tarmali and Ms. Taniya Basu
Ms. Indu Tarmali and Ms. Taniya Basu are Third-Year B.A. LL.B. (Hons.) students at The West Bengal National University of Judicial Sciences, Kolkata
Leave a Reply