Mind Over Machine - Reimagining Indian IP Law In The Age Of AI
Author: LawSphereX
Category: General Law
Published:
Updated:
MIND OVER MACHINE: REIMAGINING INDIAN IP LAW IN
THE AGE OF AI
Dhruv Verma, Vivekananda Institute of Professional Studies
ABSTRACT
Artificial intelligence (AI) is bringing about a huge change in creativity and innovation, generating works like literature, art, music, and inventions with minimal human input. This poses a direct threat to the Indian IP framework which is based on much older statutes which can’t cope up with the changes happening in the IP world. The ambiguity around AI-generated works like defining the “author” for copyright or “inventor” for patents—creates a legal vacuum that risks stifling innovation or either devaluing human creativity. The use of copyrighted material in AI training further complicates matters, as seen in cases like ANI Media v. OpenAI.
This paper investigates these challenges, focusing on Indian copyright law’s application to AI-generated content and patent law’s treatment of AI-driven inventions, by analysing India’s legal provisions, key cases like the “Raghav” copyright registration and “DABUS” patent applications, and international approaches in the UK, U. S. , EU, and WIPO discussions. The fact that India needs clearer rules on human contribution and possibly a sui generis right for AI outputs will become more and more clear as one goes through this research paper and also recollect information about day-to-day events involving IPR. The ANI Media case underscores the urgency of addressing training data issues.
Recommendations include amending existing laws to define human roles, regulating training data use, and exploring new rights to balance innovation with fairness, ensuring India’s IP regime remains robust and competitive in the global AI landscape.
Introduction
The way Artificial Intelligence has been becoming more and more mainstream ever since ChatGPT became famous is a case that would be talked about when AI finally becomes the de facto ruler of human conscience and spirit. The recent Ghibli case where AI could generate images of people in Ghibli style which is a Japanese animation studio which uses handcrafted frames and a lot of human effort shows how much of a purloiner AI is. As a concept it has existed since quite a while but it has also gained consciousness recently when a large number of new innovations in the AI field have taken place. Tools like ChatGPT, Grok, Gemini, DeepSeek, Midjourney, DALL-E, and Stable Diffusion are crafting novels, paintings, symphonies, code, and even scientific breakthroughs that rival or surpass human efforts.
This isn’t a outbreak in the technological world, it is like a double-edged sword and it also has it’s problematic side which is the problem it poses for the Indian Legal IP framework. The Copyright Act, 1957, protects literary, artistic, and musical works, while the Patents Act, 1970, covers novel inventions, both assuming a human mind drives the process. But when an AI generates a full-length novel from a single prompt or proposes a new drug formula, the question(s) that arises are who gets to claim ownership? The user who typed a few words? The developer who built the AI? Or does the work belong to no one, free for anyone to use? These questions are fuelling intense debates among legal scholars, policymakers, tech companies, and creators, and India’s IP framework is struggling to find answers.
Generative AI operates by learning from massive datasets—books, images, songs, code, often copyrighted—to identify patterns and produce new content. A prompt like “write a historical drama about a story about a girl who became the pirate queen” can yield a detailed manuscript, while AI-driven drug discovery can identify novel compounds faster than human researchers. This blurs the line between tool and creator, raising two core issues: the legality of using copyrighted material for AI training and the ownership of AI-generated outputs. In India, cases like the “Raghav” copyright registration, where an AI was briefly listed as a co-author, and the “DABUS” patent applications, which named an AI as the inventor, expose the cracks in our laws.
The Copyright Office and Patent Office have stuck to human-centric rules, but this leaves AI-generated works in a legal limbo, potentially discouraging investment in AI or flooding the public domain with unprotected content. This is not just a battle between Humans and AI, it is a battle between the rich and the poor too. Who are the people developing AI? The rich and who are the people who suffer the most due to AI? It is the middle class and the downtrodden. AI is revolutionizing sectors, from entertainment to pharmaceuticals, and India, with its expanding tech industry, cannot afford to fall behind. If works created by AI lack protection, companies may be reluctant to advance AI technology, worried that their results will be easily reproduced.
However, providing IP rights too readily—for instance, for each prompt-generated output—might diminish the worth of human creativity, enabling anyone using an AI tool to assert copyright or patents with little effort. The current ANI Media v. OpenAI case in the Delhi High Court, which examines the use of copyrighted news content for AI training, indicates that courts are beginning to address these matters, possibly advocating for more definitive regulations. This paper dives deep into how India’s IP laws handle AI-generated works, with a primary focus on copyright and a secondary look at patents. I’ll dissect the Copyright Act’s provisions, like Section 2(d)(vi) on “computer-generated works,” and the Patents Act’s human-only inventor rule, highlighting their limitations in the AI era.
Key cases like “Raghav,” “DABUS,” and ANI Media—will show how India’s legal system is navigating these challenges. I’ll also explore how jurisdictions like the UK, U. S. , EU, and WIPO discussions are tackling AI and IP, looking for lessons India can adapt. The ultimate goal is to propose a path forward that keeps India’s IP laws relevant as well as accommodate AI innovation and respect human creators. By combining statutory analysis, case studies, international perspectives, and policy evaluation, this paper aims to unravel the legal chaos and suggest practical reforms for India in the age of AI.
Methodology
To address this complex issue, I’m digging into India’s IP laws to see how they apply to AIgenerated works and inventions. The primary focus is the Copyright Act, 1957, which governs literary, artistic, musical, and other creative works, with a secondary look at the Patents Act, 1970, for AI-driven inventions. I’ll analyse key provisions—Section 2(d)(vi), Section 13, and Section 17 for copyright, and Section 2(1)(j), Section 3(k), and Section 6 for patents—to identify their human-centric biases and gaps when dealing with AI. This involves examining the statutory language, legislative history, and judicial interpretations, particularly how courts define “author,” “originality,” and “inventor” in the context of technology. For example, I’ll look at how the “modicum of creativity” standard in copyright law, shaped by cases like Eastern Book Company v. D. B.
Modak (2008), applies to AI outputs that lack human intent. I’ll also analyse landmark cases to understand how India’s legal system is able to cope up with AI. The “Raghav” case, where an AI was briefly recognized as a co-author before the Copyright Office reversed its decision, and the “DABUS” patent applications, rejected for naming an AI as the inventor, are critical to gauging current judicial and administrative thinking. The ongoing ANI Media v. OpenAI case in the Delhi High Court, which addresses the use of copyrighted news content for AI training, will provide insight into emerging judicial trends on data usage and IP rights. These cases highlight the practical challenges of applying human-centric laws to AI and reveal inconsistencies in administrative decisions.
We cannot study such cases in vacuum too so to broaden the perspective, I’ll compare India’s approach to those of the UK, U. S. , and EU, which are navigating similar issues. The UK’s Copyright, Designs and Patents Act, 1988, offers a model for “computer-generated works,” while the U. S. ’s strict human authorship rule and the EU’s “intellectual creation” standard provide contrasting approaches. Reviewing discussions at the World Intellectual Property Organization (WIPO) to understand global trends and the need for cross-border harmonization and finally evaluating policy options—amending existing laws, creating a sui generis right for AI outputs, or relying on contractual solutions like AI platforms’ Terms of Service—to propose a framework that balances AI innovation with IP principles.
This approach combines detailed statutory analysis, in-depth case studies, international comparisons, and comprehensive policy evaluation to suggest how India can adapt its IP laws for the AI era while preserving the value of human creativity. Hypothesis I argue that India’s current IP framework is inadequate for handling AI-generated works and inventions because it relies on human authorship and inventorship, concepts that don’t easily apply to AI’s capabilities. The Copyright Act’s Section 2(d)(vi), which defines the “author” of a “computer-generated work” as the person who “causes” it to be created, is too ambiguous for generative AI, where a single prompt—like “paint a surreal forest”—can produce a complex work with minimal human effort.
The requirement of “originality,” tied to human skill, labour, and judgment, is problematic for AI, which lacks consciousness or creative intent, even if its outputs appear highly original. For patents, the Patents Act’s human-only “true and first inventor” rule under Section 6 fails to account for AI’s growing role in generating novel solutions, such as drug formulas or technical designs, where human input might be limited to providing data or setting parameters. This creates a “protection gap” where AI-generated works or inventions could fall into the public domain, discouraging companies from investing in advanced AI systems due to the lack of IP protection. The hypothesis that forms here is that bringing about an amendment in the Copyright Act to define ‘human contribution’ as requiring a set of significant human input and creativity would provide clarity for copyright protection.
For patents, maintaining the human inventor rule but explicitly recognizing AI as a tool, not the inventor, would align with current law while accommodating AI’s role in innovation. In the long term, a sui generis right for AI-generated works, with tailored protection terms (e. g. , shorter durations), could address fully autonomous outputs, ensuring innovation is incentivized without distorting traditional IP principles or flooding the system with low-effort claims. Artificial intelligence, particularly generative AI, is transforming how we create and innovate and that might not sound that concerning at the first glance but if we dive deeper it’s a whole different picture.
Unlike traditional software that follow a strict set of human instructions, generative AI—like ChatGPT, Grok, Gemini, DeepSeek, Midjourney, DALL-E, and Stable Diffusion—produce new content by learning from vast datasets. These datasets include books, images, music, code, and scientific papers, often copyrighted, which the AI analyses to identify patterns, styles, and structures. For example, an AI trained on thousands of classical compositions can generate a symphony when prompted with “compose music in Mozart’s style. ” The output isn’t a direct copy but a novel synthesis, raising two critical legal issues: the use of copyrighted material in training and the ownership of the generated work. The training phase is a legal minefield.
The internet becomes a junkyard for these AI models to use datasets and all types of information to create these outputs, including copyrighted novels, articles, images, or songs. This has given rise to global lawsuits, such as the ANI Media v. OpenAI case in India’s Delhi High Court, where news outlets claim that their content was used to train AI without permission or compensation. If courts rule this constitutes infringement, AI developers might need to secure licenses for training data, increasing costs but ensuring fairness to creators. This issue is critical in India, where media and creative industries are vibrant, and unauthorized use could harm content creators. A ruling in ANI Media could set a precedent for compulsory licensing or royalties, reshaping how AI companies operate in India. The generation phase is equally complex.
Human involvement in AI creation varies along a spectrum. At one end, AI is a tool, like Photoshop’s AI filters, where the human retains full creative control. At the other end, AI can produce works with minimal input, like a single prompt (“write a thriller novel”), or theoretically create autonomously without any human nudge or interference. Most cases fall in the middle, where users provide prompts and refine outputs to varying degrees. This spectrum challenges IP law, which assumes a human creator with clear intent and effort. If a user’s role is limited to a prompt, is it enough to claim authorship or inventorship? The answer depends on how we define “creation” in the AI era India’s Outdated Copyright Law India’s Copyright Act, 1957, is firmly human-centric.
Section 13 lists copyrightable works— literary, dramatic, musical, artistic, films, and sound recordings—but assumes a human author. Section 17 states that the “author” is the first owner of copyright, except in cases like workfor-hire or commissioned works. The Act defines “author” differently for each work type (e. g. , the composer for music, the artist for paintings), but it’s always a natural person. Section 2(d)(vi) offers a partial solution, defining the “author” of a “computer-generated work” as the person who “causes” it to be created. Introduced in 1994, this provision was designed for simpler technologies, like CAD software or basic algorithms, where a human directly controls the output. But with generative AI, where a prompt like “create a Van Gogh-style painting” can produce a masterpiece, the “causes” clause feels outdated.
Does typing a one-line prompt qualify as “causing” the work, or does it require more, like iterative refinement or creative editing? Originality is a major hurdle. Indian courts, following cases like Eastern Book Company v. D. B. Modak (2008), define originality as requiring “skill, labour, and judgment” or a “modicum of creativity. ” This standard assumes human intellectual effort and intent. An AI might be a able to create outputs based on human inputs and human generated responses but in no way can it match the intellect of a real human, i. e. a human brain. For example, an AI-generated poem might rival Shakespeare or Willaim Wordsworth, but without human intent, it might fail the originality test, falling into the public domain.
This creates a dilemma for companies to develop AI for creative purposes, as their outputs might be unprotected and freely usable by competitors or the public. Some companies might actually want to do a little good to the industry as their AI model can help small businesses grow but they get overshadowed by these bigger companies which often use models that produce a cheaper output. The “Raghav” case is a stark example of this tension. In 2020, artist Ankit Sahni sought copyright for a painting, “Suryast,” created by an AI tool called Raghav, listing the AI as a co-author. The Indian Copyright Office initially granted the registration, a surprising move that suggested openness to AI authorship. But it later cancelled the registration, likely because the Copyright Act doesn’t recognize non-human authors.
While not a judicial precedent, this reversal highlights the practical challenges of applying a human-centric law to AI. If Sahni’s minimal prompt was enough to claim authorship, it risks lowering the bar for originality, allowing anyone with an AI tool to claim copyright for trivial inputs. But denying protection altogether could discourage investment in AI tools like Raghav, as developers and users lose the ability to control their outputs. Other similar theories for copyright attribution don’t fully resolve the issue. One approach is to grant authorship to the AI’s developer, who designed and trained the model. But developers often have little direct input in specific outputs, and their role feels too remote to meet the “skill and judgment” standard. For example, OpenAI’s developers don’t directly craft every ChatGPT output; the model generates based on learned patterns.
Another theory is to treat the user who provides the prompt as the author, as many AI platforms’ Terms of Service (e. g. , OpenAI, Midjourney, DeepSeek, Gemini, Grok) assign rights to users. However, a simple prompt like “draw a cat” offers minimal creative direction, and granting copyright for it could trivialize originality, the problem that arises is to selectively find out where the concept of originality is being hampered with. Joint authorship, where a human and AI collaborate, is plausible when the human heavily edits the output, like refining an AIgenerated story into a polished novel. But it falters for prompt-driven or autonomous works where human contribution is negligible. If no human author can be identified, the work might default to the public domain, creating a “free for all” scenario that undermines incentives for AI development.
Patent Law: A Similar Picture What is a patent? It is a type of Intellectual Property where the owner gets an exclusive right over its usage. Patents present similar challenges. The Patents Act, 1970, protects inventions— new, non-obvious, industrially applicable solutions—under Section 2(1)(j). Section 6 restricts patent applications to the “true and first inventor” or their assignee, explicitly meaning a human. AI-driven inventions, like a novel drug formula or technical design, complicate this. If an AI analyses data and proposes a solution, the human’s role—perhaps just feeding data or setting parameters—might not qualify as “invention. ” The “DABUS” case is a prime example. In 2019, a famous computer scientist named Stephen Thaler applied for patents in India, naming his AI, DABUS, as the inventor of a food container and a “neural flame.
” The Indian Patent Office rejected the applications, citing the human-only inventor rule, aligning with decisions in the U. S. and EU. South Africa granted the patent, but this was an administrative decision without rigorous legal scrutiny, offering little guidance for India. Section 3(k) adds another hurdle, excluding “computer programmes per se” from patentability. While AI algorithms themselves might not qualify, inventions with a “technical effect” (e. g. , a new device or process enabled by AI) could. For example, AI systems like IBM’s Watson have identified novel compounds for medical use, but attributing inventorship to a human who merely provided data stretches the definition of “inventor” thin, this would be a disrespect to actual human inventors who put their heart and soul in their patents and designs.
This gap could kick-start research into AI-driven innovation, particularly in areas like pharmaceuticals or engineering, where AI can accelerate discoveries but humans need credit to secure patents. If the AI autonomously identifies the inventive step, the human’s role feels too small to claim inventorship, leaving a lacuna or grey area. A Global Perspective: Lessons for India Other jurisdictions offer valuable insights for India. The UK’s Copyright, Designs and Patents Act, 1988, is one of the few laws addressing “computer-generated works. ” Section 178 assigns authorship to the person who makes the “arrangements” for the work, which could be the programmer, user, or computer owner. This flexibility is appealing but vague for modern AI, where the “arranger” isn’t always clear.
For example, if a user prompts Midjourney to create an artwork, is the user or the platform’s developer the author? The U. S. takes a stricter stance: the Copyright Office’s 2023 guidance requires human authorship, denying copyright to AIgenerated works, as seen in the Zarya of the Dawn comic book case, where AI-made images were excluded but human-written text was protected. The U. S. Patent Office similarly demands human inventors, rejecting DABUS. The EU ties copyright to “intellectual creation,” implying human effort, and also rejected DABUS as an inventor. A common consensus cannot be reached regarding the legitimacy of AI generated patents or where AI is involved even remotely. The European Patent Office argued that only natural persons can invent, emphasizing the human-centric nature of patent law.
The World Intellectual Property Organization (WIPO) is facilitating global discussions, but no unified framework or statute has emerged. WIPO’s Conversation on IP and AI highlights the need for international cooperation, as AI content crosses borders, creating potential conflicts between jurisdictions with different rules. For instance, an AI-generated artwork protected in the UK might be public domain in the U. S. , complicating global distribution. Indian government claimed in 2024 that our existing laws are sufficient, arguing that human input, even minimal, can assign authorship or inventorship. But this feels like a temporary fix. The ANI Media v. OpenAI case in the Delhi High Court, where news outlets challenge the use of their content to train AI, suggests courts are starting to address these issues.
This case is crucial as a ruling here could the answer the question whether training on copyrighted material requires licensing, potentially setting a precedent for balancing creator rights with AI development. India must consider these global and domestic developments as soon as possible to avoid falling behind in the AI race.
India’s IP law framework creates a difficult scenario where you cannot benefit by sacrificing something or the other. If AI-generated works or inventions receive no protection, companies might hesitate to invest in AI, which might slow down innovation in fields like art, music, or biotech. For example, a company developing an AI that autonomously creates marketable artworks might shift investment to jurisdictions with clearer laws if India offers no protection. Conversely, granting copyright or patents too easily—say, for every prompt-driven output— could flood the system with low-effort claims, devaluing human creativity. The “Raghav” cancellation and “DABUS” rejections show India’s adherence to traditional rules, but this approach doesn’t scale for AI’s growing autonomy. Training data is another flashpoint.
AI models rely on vast datasets, often including copyrighted material, raising infringement risks. The ANI Media v. OpenAI case could set a precedent for requiring licenses or compensation, which would protect creators but increase costs for AI developers. India needs a balanced approach—allowing data access for innovation while ensuring fairness. Cross-border issues add complexity: AI content flows globally, so India’s lack of clarity could spark disputes with jurisdictions like the EU, which emphasizes creator rights. Without reform, India risks losing its edge in the global AI market, where countries like the U. S. and EU are already shaping policy through debates and rulings. Conclusion and Recommendations AI’s ability to create and invent is exposing critical flaws in India’s IP framework.
The Copyright Act, 1957, and Patents Act, 1970, are built for human creators, leaving confusion over AI-generated works and inventions. The “Raghav” case, where an AI co-authorship claim was reversed, and the “DABUS” patent rejections, which denied AI inventorship, show the system’s struggling to adapt. The government’s 2024 stance that no new laws are needed feels shortsighted when AI can produce complex works with minimal human input. This uncertainty risks discouraging AI investment or undermining human creativity by granting rights too loosely. The ANI Media v. OpenAI case, addressing copyrighted training data, underscores the urgency of reform, as courts may soon demand clearer rules. To address this, India should adopt a multi-pronged strategy: • Amend Copyright Law: When the law becomes outdated, we amend it.
A similar approach can be undertaken in IPR as the statutes related to IP law in India were formulated in the 20th century only. For instance amending Section 2(d)(iv) to define ‘causing’ as a work requiring significant human effort such as editing, refining or shaping AI output. This ensures that only human-driven works get copyrighted preserving the originality standard. • Clarify Patent Rules: Maintaining the human inventor rule in the Patents Act while at the same time explicitly recognizing AI as a tool, not the inventor. Focus on the human’s role in defining the problem or interpreting AI output to secure inventorship. For instance, a researcher using AI to identify a new drug compound should be the inventor if they framed the research question or validated the result, ensuring patent law accommodates AI-driven innovation too.
• Regulate Training Data: Establish clear guidelines for using copyrighted material in AI training, ensuring creators are compensated fairly, possibly through licensing models or royalties, AI should not become a credit stealer at any point. The ANI Media v. OpenAI case could provide judicial guidance, potentially requiring AI developers to secure permissions, balancing creator rights with innovation needs. This could involve a statutory framework for data licensing, similar to music royalty systems. • Explore Sui Generis Rights: In the long term, consider a sui generis right for AIgenerated works, offering shorter protection periods (e. g. , 10–15 years) for fully autonomous outputs. This stimulates innovation without overwhelming the traditional IP system or diluting human creativity.
For example, an AI-generated artwork could receive limited protection, allowing creators to profit while preventing monopolies over machine outputs. • Promote Education and Awareness: India is still a backward country when it comes to awareness about new technology and innovation in general. Many parts of India are still facing problems like poverty and illiteracy. Therefore the government should take the initiative of making people aware about the growth of concepts like AI so that we all can collectively give our inputs relating to the laws and ordinances affecting the use of AI and innovation. India possesses a distinctive chance to take the lead in this developing area, harmonizing AI's transformative capabilities with the tenets of IP law.
India can enhance its IP system to benefit both human creators and AI advancement by promptly implementing specific reforms including revising current laws, managing training data, and investigating new rights. Not adapting risks lagging behind global fronts like the U. S. or EU, where policy discussions and court decisions are already influencing the future of AI and IP. Cases such as ANI Media v. OpenAI are urging courts to respond, making it essential for India to develop a progressive framework that utilizes AI’s capabilities while safeguarding human creativity.
References • Copyright Act, 1957, Sections 2(d)(vi), 13, 17 (India). • Patents Act, 1970, Sections 2(1)(j), 3(k), 6 (India). • Eastern Book Company v. D.B. Modak, AIR 2008 SC 809 (India). • “Raghav” Copyright Registration Case, Indian Copyright Office (2020). • “DABUS” Patent Applications, Indian Patent Office (2019). • ANI Media (P) Ltd. v. Open AI Inc., Delhi High Court (ongoing as of 2024). • UK Copyright, Designs and Patents Act, 1988, Section 178. • U.S. Copyright Office, Guidance on AI-Generated Works (2023). • World Intellectual Property Organization (WIPO), Conversation on IP and AI (ongoing discussions).