Abstract
Generative artificial intelligence (AI) technologies like ChatGPT, MidJourney, and DALL•E have changed how we think about creativity and copyright. According to the Indian Copyright Act of 1957, only humans can be seen as authors, leaving works created by AI in a confusing legal situation. Around the world, people are discussing whether copyright should belong to the AI’s developer, the user, or become public property. This paper offers a new idea: considering prompts as authorship. It claims that prompts—made by human users to direct AI outputs—can be seen as a type of creative expression, meeting the “skill and judgment” standard set by the Supreme Court in Eastern Book Company v. D.B. Modak[1].
By viewing prompts as a key part of human creativity, Indian copyright law can fix the ownership issue around AI-assisted works without giving authorship to machines. This system keeps human control while encouraging new ideas and stopping AI developers from having too much power. The study looks at other countries like the UK, which acknowledges “computer-generated works” in Section 9(3) of the Copyright, Designs and Patents Act, 1988[2], and the United States, which insists on human authorship (Zarya of the Dawn, 2023). It suggests possible changes for India. The analysis shows that recognizing prompt-authorship can help protect creators, support innovation, and keep the legal framework clear. The paper ends by suggesting that the courts or lawmakers should clearly recognize prompt-authorship as a special category in Indian copyright law.
Keywords: Generative AI, Copyright Law, Authorship, Prompt Authorship, Indian Copyright Act, 1957, AI-Generated Works.
Introduction
Copyright legislation seeks to safeguard original human thoughts and works. Under the law of India, through its 1957 Copyright Act[3], an “author” is limited to a human being only, conferring rights on human beings only when they create literary, artistic, dramatic, and musical works. But this equation has altered now with new generative AI technology. Software such as ChatGPT, MidJourney, and Stable Diffusion enable users to write poems, songs, paintings and even movie scripts just by inputting prompts into a keyboard. Such works sometimes equal or even supersede human creativity, yet pose a critical question of law: ownership of copyright by whom, if anybody?
Classic arguments have centered on three positions. First, some argue that developers of AI or platform owners should be authors because construction of the model and training data make output feasible. Second, others believe that users of AI should be authors because users give specific prompts that determine output. A third proposal denies copyright altogether and advocates that AI creations should automatically be common property because no human creativity is going into creations. Each of these proposals has flaws: having developers as authors can mean that tech firms have undue rights, user authors are unclear as a matter of present law, and placing AI output in public domain can discourage investment in creativity and business.
This paper suggests a new model—prompt as authorship. Instead of seeing AI as an independent creator, we focus on the human role in making prompts. Prompts are not just simple instructions; they are creative ideas that influence the uniqueness, quality, and path of AI results. For example, the difference between a basic question (“write a poem on love”) and a detailed prompt (“compose a Shakespearean sonnet about unrequited love in the digital age”) shows how human ability and choice affect the final piece.
Indian copyright law says that “minimal creativity” is enough for protection. The Supreme Court in the case Eastern Book Company v. D.B. Modak (2008)[4] used the “skill and judgment” test. This means a work needs to show some effort and thought, but not necessarily great artistic talent. Following this idea, prompts that are original enough can be seen as “literary works” under Section 2(y) of the Copyright Act[5].
The principal argument of this paper is that ownership of copyright should reside in the owner of the prompt and not in the AI or its creator. This paper studies a variety of different legal, comparative, and policy positions to show that prompt-authorship can be a suitable set of legal solutions for India.
Review of Literature
Indian Position:
Indian copyright law has not yet dealt with works created by AI. Section 2(d) of the 1957 Copyright Act only defined “author” in human terms. For example, a book author is “the author of the work” and in films and in sound recordings is given as author being a producer. This further suggests that it is not possible to determine a non-human author like AI. As compared to UK’s Copyright, Designs and Patents Act of 1988 (Section 9(3)), India has no statute dealing with works created by computers.
One of the principal Indian cases is Eastern Book Company v. D.B. Modak (2008), in which the Supreme Court did not embrace the “sweat of the brow” approach and rather employed the “skill and judgment” threshold of originality. According to the Court, creativity requires actual thinking rather than mechanical labor. This ruling has a direct impact on this present work: stimuli prepared with innovative intention could attain this threshold of originality as a demarcation from mere or typical instructions.
In academic writing, Dr. N.S. Gopalakrishnan and T.G. Agitha’s book Principles of Intellectual Property (2014)[6] talks about the importance of human authorship in Indian copyright law but mentions the growing conflict caused by digital technologies. Although they did not predict generative AI directly, their analysis highlights the strictness of the current Act, which this research claims needs to change.
Global Scholarship
In America, the Copyright Office will invariably require human authorship. An important ruling occurred in the “Zarya of the Dawn” (2023)[7] case wherein protection did not accrue to images created by AI through MidJourney, although a story authored by a human ratified. This indicates that AI outputs in the opinion of the US cannot attain copyright. Such scholars as Pamela Samuelson (Artificial Intelligence and Copyright, 2019)[8] endorse this view and advise against diluting human creativity, which is copyright’s foundational principle.
In the UK, Section 9(3) of the Copyright, Designs and Patents Act, 1988[9] says that “the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken” for computer-generated works. This legal rule looks ahead but has not been tested in court during the time of generative AI. Experts like Lionel Bently and Brad Sherman in Intellectual Property Law (5th ed., 2018)[10] believe that Section 9(3) is a practical agreement but is unclear when it comes to creative prompts, leaving it uncertain if the “arranger” is the programmer, platform owner, or end-user.
In Thaler v. Comptroller-General of Patents (UK, 2021)[11], the Court of Appeal ruled that AI (DABUS system) could not be an “inventor” within patent law. This is a patent law case but reflects that courts do not want to countenance AI as a legal person or as a person capable of having rights in intellectual property laws.
Gap in Literature
Comparative analyses most often frame the debate as a binary question: should rights be of the AI maker or of its user? Even prospective discussions (such as Tanya Aplin’s work Copyright and Artificial Intelligence, 2021)[12] do not treat the prompt as a creative work. Little attention has been paid to developing skill of prompt engineering that is gaining a reputation in practice as a gifted creative field that decides outputs in such fields as marketing communications, digital arts creation, and music production.
This paper shall attempt to harmonize this subject of prompt-authorship with Indian copyright law. It will look into it through the prism of the standard of originality set in D.B. Modak[13] and shall keep in view positions from the US and UK.
Research Methodology
This work follows a doctrinal approach to law, informed by comparative and policy analysis.
1. Statutory Analysis: A close reading of the Copyright Act, 1957, particularly Sections 2(d) (author), 2(y) (literary work), and 13 (works in which copyright subsists)[14]. The study also examines the Information Technology Act, 2000[15], for any overlaps regarding electronic records.
2. Case-Law Analysis: Eastern Book Company v. D.B. Modak (2008, on originality) is one of such cases as is University of London Press v. University Tutorial Press (1916, on test of originality) and some others including Thaler v. Comptroller-General of Patents (UK, 2021) and Zarya of the Dawn (US Copyright Office, 2023).
3. Comparative Legal Study: Focused analysis of UK’s Section 9(3) of the CDPA[16], US Copyright Office policy, and EU’s evolving AI Act to evaluate possible models for India.
4. Policy Analysis: Considering what identifying prompt-authorship implies for creative businesses, tech firms, and end-users in India.
This integration allows the paper to connect diverse legal principles, frame Indian law within an international comparison, and suggest plausible amendments suitable to India’s social and economic milieu.
Problem 1: Originality of Prompts as Provided in Indian Copyright Law
The central question of debate surrounding ownership of prompts is if prompts are at all original so as to fulfill thresholds of Indian copyright law. Unless such prompts are deemed to be original work pieces, all arguments of ownership of copyright of AI-assisted outputs stand incorrect.
The Legal Standard of Originality in India
The Indian Copyright Act, 1957[17], does not explain “originality.” Instead, Indian courts have looked at the idea through past cases. At first, Indian law focused on the “sweat of the brow” idea, where just putting in work was enough to show originality. But this changed a lot when the Supreme Court made a ruling in Eastern Book Company v. D.B. Modak (2008)[18]. The Court decided that originality needs “a minimum degree of creativity” shown through skill and judgment that is not just mechanical. Importantly, the Court explained that originality does not need to be new or genius but requires a creative touch that sets the work apart from ordinary or unimportant efforts.
This test reflects the Canadian method used in CCH Canadian Ltd. v. Law Society of Upper Canada (2004). It turned down both the “sweat of the brow” test and the “modicum of creativity” test. Instead, it chose a balanced way that focuses on intellectual effort.
Originality in India is found if work reveals human thinking and innovative decision-making as against being a product of routine or ordinary work.
Answer:
The main point of the analysis is whether a prompt—a written instruction given to an AI system—can meet this standard. At first look, prompts might seem unimportant, made up of short phrases (for example, “draw a cat wearing sunglasses”). These simple instructions are not likely to be considered original works. But things are different when prompts are made with careful creativity, detail, and imaginative ideas.
For instance, a question such as “Make an impressionist painting of the Varanasi ghats at dawn, with saffron hues casting a reflection on the Ganges, as a tribute to Claude Monet” indicates a blending of creative decisions—style, subject, location, and artistic inspiration. This requires the talent and discretion of the user to think of and articulate the creative thought. Such questions demonstrate the thought effort and creative drive required under D.B. Modak.
Thus, such prompts can be grouped along with literary works of Section 2(y), which embraces “tables and compilations” as well as is not limited to voluminous texts. Shortness of such prompts is no indication that they are Not Protectable; Indian courts have opined that short creative works such as slogans or titles can be registered provided such works demonstrate creativity (for instance, in Pepsi Co. Inc. v. Hindustan Coca Cola Ltd. (2003) a slogan that is protectable)[19].
Doctrinal Matters
There are two doctrinal difficulties.
1. Fixation: Work must be “fixed” in some physical medium to come within copyright. Prompts are printed to paper when keyed in or written to disk, which meets this condition.
2. Output vs. Input: Some courts would argue that creativity is in what the AI generates rather than in the brief prompt. This ignores the fact that output is conditioned upon what is in the prompt. Taking submissions as original works respects human effort behind the AI and is consistent with copyright’s aim.
Comparative Insight
In America, the Zarya of the Dawn ruling of the Copyright Office (2023)[20] agreed that a human author provides text but did not incorporate pictures produced by an AI. Similarly, India might also view prompts as human-written texts during creative work and differentiate those from mechanical work of an AI.
The UK’s Section 9(3) of the CDPA provides credit to computer-created works to the arranger of the work. This can also be interpreted as endorsing prompt-authorship. Under this idea, prompts are the “arrangements necessary” to make the work.
Conclusion on Originality
In this skill and judgment test, more than mere instructions and creative human imagination can be regarded as original literary pieces according to Indian law. This consideration provides a basis in law to regard the prompt engineer as a creator of AI-assisted works. Not all will make it through, but courts can examine each situation separately to ensure that only genuinely creative inputs receive copyright protection and not mere commands.
Thus, having distinctive prompts is as much in conformity to present laws as it is necessary to preserve human control in the age of AI.
Problem 2: Authorship and Ownership of AI-Generated Compositions
The second big challenge is determining who must be deemed to be the author and copyright owner of AI-produced works. This is no trivial matter but rather a question of striking a core balance between creativity, technology, and economic compensation in the age of information.
The Indian Legal Framework
In Section 2(d) of the Copyright Act, 1957[21], only human beings can be authors:
- For books and art, the author is the person who makes the work.
- For films and recordings, such as those of sounds, the producer is himself an author.
The Act does not contemplate authorship by agents that are not human like AI. Therefore, AI cannot have copyright ownership as per Indian law. So, is ownership of works generated by AI residing in the prompt engineer (user) or AI system constructor or should such works be in the public domain?
Option 1: Prompt Engineer as Author
The optimal solution, which is espoused in this essay, is that the individual creating the prompts is to be viewed as the author. This aligns with original work analysis in Issue 1—if literature prompts are original works, then authorship is to reside with the individual that infuses that originality.
In the case Pepsi Co. Inc. v. Hindustan Coca Cola Ltd. (2003)[22], the Delhi High Court accepted that slogans can be original, showing that being short does not mean they lack authorship. Similarly, creative prompts are a human input that allows for copyright protection. Seeing the user as the author also supports the constitutional aim of protecting personal creativity under Article 19(1)(g) (the right to trade or work) and encouraging artistic industries.
Option 2: Software/Platform as the Author
Another approach is to accord authorship to that developer of the AI system. This is premised upon that nothing would be generated without the designer and training data of the developer. This principle is comparable to that found in Section 9(3) of the UK CDPA 1988 that provides that authorship of works generated by computers belongs to that person responsible for having made “arrangements necessary” for creation.
In a generative AI world, this would mean a technological monopoly. When all content of MidJourney or ChatGPT is that of a platform, users are less motivated to create, and small producers are excluded. It also mistakenly assigns credit: developers create the tool, the content itself. Courts in a situation such as Eastern Book Company v. D.B. Modak show human work to be determinative of originality—something developers do not bring to bear upon creation.
Option 3: Public Domain Method
Another perspective is that AI works must be excluded from copyright and placed directly in the public domain. This practice is exemplified in the US Copyright Office decision in Zarya of the Dawn (2023) when it refused to grant copyright to AI paintings. This benefit is avoiding monopolization and keeping it free to the general public.
But the public domain approach could hurt creative industries. If AI-assisted works are not protected, commercial creators might not want to spend time and effort on them, which could hurt the economy. Also, it unfairly ignores the human ideas behind the prompts.
Balancing User Agreements and Copyright Law:
Any complicating factor is that AI platforms often have Terms of Service that determine who owns what. For example, OpenAI allows users to have rights over what they create but keeps a license for itself; MidJourney, on the other hand, gives more commercial rights to its subscribers. These agreements are legally binding, but they cannot change copyright laws. The Copyright Act should be the main rule when solving ownership problems.
Conclusion on Authorship and Ownership:
Authorship is-
The most suitable and justifiable stance is to treat the prompt engineer as the writer of work produced by AI aid. This stance:
- Secures human agency, in accordance with Indian jurisprudence.
- Prevents monopolization of developers.
- Provides financial compensation to authors.
- Can exist along with licensing agreements.
Indian copyright law can be made to accommodate generative AI realities as a result of human input without diluting its human-oriented emphasis.
Problem 3: Missing Policies and Laws in India
The third question is whether India’s existing system of copyright is sufficient in AI-created works, or whether we must alter laws to cope with issues resulting from generative AI. Over time, the 1957 Copyright Act has been revised to incorporate digital technology but remains centered upon creativity by human beings. Generative AI defies this notion significantly and reveals gaps within a law that must be corrected.
The Current Situation:
The statute refers to “author” (Section 2(d)) only as natural authors like composers, artists, and writers. No section like Section 9(3) of the UK Copyright, Designs and Patents Act, 1988 that deals explicitly about computer-produced works exists. This gap leads to Indian courts having no express directive as to how to approach outputs from AI.
Section 13 of the Act states that copyright subsists in “original” literary, dramatic, musical, and artistic works. Originality is the test that must be passed by such works as interpreted in Eastern Book Company v. D.B. Modak (2008)[23], which calls for “skill and judgment.” Whether AI-assisted creations meet this qualification is doubtful, and doubt generated by ambiguous laws is harmful to creators, developers, and businesses.
Comparative Models
1. United Kingdom (UK CDPA, 1988, Section 9(3))
UK approach grants rights of computer-generated work to the person who agreed to have them produced. This is simple but might have a bias towards developers or companies over end-users.
2. United States (Human Authorship Doctrine)
Human authorship is required by the U.S. Copyright Office, as reaffirmed in Zarya of the Dawn (2023). Outputs from AI would be excluded from protection if they are all machine-made. This approach prevents corporate monopolies but might leave worthwhile work without protection.
3. European Union (EU AI Act, 2024)[24]
The EU does not give copyright to AI outputs yet, but it does have rules for AI transparency and accountability. This might influence copyright by making people say if they used AI in creative work.
India’s Conundrum:
India is at a policy juncture. Failure to protect AI-created work may facilitate piracy, lower incentives to create in creative sectors, and drive Indian content creators to foreign platforms that have clearer rights frameworks. Giving developers ownership rights may allow corporate monopolization to displace individual creators and cultural sectors.
The middle path—acknowledging who creates prompts—needs changes in the law. By clearly adding AI-assisted works and giving authorship to the prompt engineer, India can provide protection for people and adapt to technology.
Conclusion
As generative artificial intelligence obscures the line between invention and reproduction, Indian law governing copyright, needs to adapt in order to retain the spirit of human imagination. Identifying prompt-authorship ensures that technology is a tool and not a substitute for human mind. It strikes a balance between innovation and responsibility, enabling creators while restraining monopolies. The future lies in re-envisioning authorship, not in dismissing it. India needs to be at the forefront of this change wherein the law keeps up with imagination and the author of the future might very well be the brain behind the machine.
Plans to transform (proposals)
1. Amending Section 2(d) [25]
Introduce a new clause defining authorship of AI-assisted or computer-generated works, modeled on but distinct from the UK’s Section 9(3), ensuring authorship vests in the human prompt engineer rather than the developer.
2. Imagining a Unique Right
India may have a limited sui generis system of AI-created works that have short copyright terms (e.g., 10 years) or restricted rights to balance protection and access to the public. 3. Define Ownership of Team AI Systems The Act must have provisions relating to joint authorship when many others assist in creating prompts or when human editing is combined with AI output. 4. Abide by the IT Act, 2000[26]. Platforms of AI are intermediaries under the IT Act. An integrated amendment might link liability, transparency, and copyright authorship to offer responsibility across regimes. Conclusion and Recommendations The development of generative AI has brought to light large issues in Indian copyright law. Pre-dating the digital age, the Copyright Act of 1957[27] assumes that works are only made by humans. It lacks the thought required to address works produced by artificial intelligence. Three of the chief difficulties identified above are: whether or not prompts are original enough, how AI creations can be regarded as having an author, and how legislation can be adapted to facilitate innovation as well as protection. The “skill and judgment” test of Indian law (Eastern Book Company v. D.B. Modak, 2008)[28] enables us to view prompts as original compositions. This consideration is favorable to ascribing authorship to prompt engineers, who are central to producing AI work. Attribution to developers would result in large tech companies dominating everything, whereas deregulating works to anyone might demotivate creators. Granting prompt-authorship is a justifiable remedy, retaining human control along with an appreciation of technology. Legislationally, India should shun two extremes: restrictive laws allowing human authors alone (like in U.S.) and overly broad laws accrediting developers (like Section 9(3) of U.K.)[29]. India can show the way by ushering a different solution that allows AI-aided creations and extends rights to human authors through prompt-authorship.
Recommendations-
1. Modify Section 2(d) of the Copyright Act to make prompt-authorship an authorship category of AI-assisted creations.
2. Imagine a sui generis right of limited duration to AI creations to strike a balanced approach between protection and public access.
3. Align copyright reform with the IT Act of 2000[30] to regulate AI platforms as intermediaries and to disperse rights in a defensible and transparent fashion. As a result of undertaking these steps, India will have a copyright regime that defends human creativity, encourages cultural industry development, and becomes a leader of global debate on AI and intellectual property.
Submitted by-
Ashish Tanmay
3rd year BA. LLB. Student
Symbiosis Law School, Nagpur
[1] Eastern Book Company v. D.B. Modak, (2008) 1 SCC 1
[2] Copyright, Designs and Patents Act, 1988, § 9(3), No. 48, Acts of Parliament, 1988 (U.K.)
[3] The Copyright Act, 1957, No. 14, Acts of Parliament, 1957 (India)
[4] Supra 1
[5] The Copyright Act, 1957, § 2(y), No. 14, Acts of Parliament, 1957 (India)
[6] N.S. GOPALAKRISHNAN ET. AL., PRINCIPLES OF INTELLECTUAL PROPERTY (Eastern Book Company 2014)
[7] Kristina Kashtanova v. United States Copyright Office, No. 23-cv-8431 (S.D.N.Y. 2023)
[8] Pamela Samuelson, Artificial Intelligence and Copyright, 62 COMMUN. ACM 20 (2019)
[9] Copyright, Designs and Patents Act, 1988, § 9(3), No. 48, Acts of Parliament, 1988 (U.K.)
[10] LIONEL BENTLY ET. AL., INTELLECTUAL PROPERTY LAW (5th ed. 2018)
[11] Thaler v. Comptroller-General of Patents, Designs and Trade Marks, [2021] EWCA Civ 1374.
[12] Tanya Aplin, Copyright and Artificial Intelligence, in INTELLECTUAL PROPERTY AND THE ROBOTIC REVOLUTION 72 (D’Agostino & Piatti eds., 2021)
[13] Supra 1
[14] The Copyright Act, 1957, §§ 2(d), 2(y) & 13, No. 14, Acts of Parliament, 1957 (India)
[15] Information Technology Act, 2000, No. 21, Acts of Parliament, 2000 (India)
[16] Supra 9
[17] The Copyright Act, 1957, No. 14, Acts of Parliament, 1957 (India)
[18] Supra 1
[19] Pepsi Co. Inc. v. Hindustan Coca Cola Ltd., 2003 (27) PTC 305 (Del)
[20] Supra 7
[21] The Copyright Act, 1957, § 2(d), No. 14, Acts of Parliament, 1957 (India)
[22] Supra 19
[23] Supra 1
[24] Artificial Intelligence Act, Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024
[25] The Copyright Act, 1957, § 2(d), No. 14, Acts of Parliament, 1957 (India)
[26] Supra 15
[27] Supra 17
[28] Supra 1
[29] Supra 9
[30] Supra 15