This Piece of Article is Authored by :- Shivangi Haldwani student at ,NLU,Delhi
Introduction
Prompt engineering has emerged as one of the most visible features of the generative AI era. In simple terms, prompt engineering refers to the drafting of instructions, questions, contextual cues, or structured inputs that are given to an artificial intelligence system in order to produce a desired output. These inputs may be brief or highly elaborate, and in some professional settings they are created with considerable skill, experimentation, and strategic thinking. As a result, a new legal question has gained importance: can an AI prompt be treated as an original literary work protected by copyright under Indian law?
This question is not merely theoretical. In India, the Copyright Act, 1957 protects original literary, dramatic, musical, and artistic works, but the law was drafted long before the rise of generative AI. The existing legal framework therefore has to be interpreted in light of modern technology. On one side, there is an argument that a carefully drafted prompt reflects intellectual effort, selection, arrangement, and creativity. On the other side, there is a strong counterargument that most prompts are functional instructions meant to control a machine, not expressive literary compositions in themselves. From an Indian perspective, this is where the major limitation lies: copyright protects expression, not mere instructions, methods, or ideas.
The Indian legal position remains unresolved in the sense that there is no specific statutory provision devoted to prompt engineering. However, the broader doctrines of originality, idea-expression distinction, human authorship, and computer-generated works offer enough guidance to assess the issue. When those doctrines are applied carefully, the conclusion becomes fairly narrow. Only a very limited category of prompts, if any, may qualify as original literary works. Most routine prompts will not.
The starting point for analysis is the Copyright Act, 1957. Section 13 of the Act provides that copyright subsists in original literary, dramatic, musical, and artistic works, cinematograph films, and sound recordings. For the present topic, the most relevant category is “literary work,” which is defined broadly in Section 2(o) to include computer programs, tables, and compilations, along with works expressed in writing, regardless of whether the writing is in printed or digital form. At first glance, this broad definition may seem to support protection for AI prompts, since prompts are written expressions.
However, copyrightability does not depend only on the fact that something is written down. The work must also be original. That requirement is crucial because the statute does not protect every written text. Section 17 of the Act deals with first ownership, and Section 2(d) defines the author. For computer-generated works, Section 2(d)(vi) states that the author is “the person who causes the work to be created.” This provision appears relevant where a human user writes a prompt and an AI system generates content. Yet the existence of this clause does not automatically mean that the prompt itself is protected. It only helps identify authorship in computer-generated works; it does not eliminate the need for originality.
The Act also excludes certain subject matter from protection through Section 52, which deals with acts not constituting infringement. More broadly, Indian copyright law is influenced by the principle that copyright does not protect ideas, methods, or functional systems. This means that even if a prompt is written in language, its legal status depends on whether it is expressive literary content or merely a functional instruction.
Originality as the Main Barrier
The most significant limitation on copyright protection for AI prompts in India is originality. Indian copyright law has moved away from a purely labour-based theory of protection. The Supreme Court’s decision in Eastern Book Company v. D.B. Modak is the leading authority on this point. In that case, the Court held that originality requires more than mere effort or sweat of the brow; there must be a minimum degree of creativity, skill, and judgment. This doctrine is central to the present issue because it means that a prompt will not be copyrightable simply because the user spent time refining it.
Applied to prompt engineering, this rule creates a high threshold. A prompt such as “write a poem about rainfall in Delhi” may be useful, but it is too simple and functional to qualify as an original literary work. Even if a prompt is long, that does not necessarily make it original. A large number of commands, modifiers, or constraints may still amount to nothing more than a detailed instruction set. Copyright law does not reward length or complexity alone. It rewards creative expression.
This is where the legal difficulty becomes clear. Most prompts are written to produce an output efficiently. They are designed for practical utility, not for self-contained literary significance. The fact that a prompt is carefully optimized does not mean it has literary value in the copyright sense. Originality under Indian law requires some independent creative expression, not just strategic drafting for machine performance. Therefore, the ordinary run of prompts will fail the test laid down in Eastern Book Company v. D.B. Modak.
Functional Language and the Idea-Expression Divide
Another major limitation is the idea-expression distinction. This is one of the most enduring principles in copyright law, and it is highly relevant to prompt engineering. Copyright protects the form in which an idea is expressed, not the idea, concept, system, or method itself. In the context of prompts, this becomes important because many prompts are essentially commands or operational instructions. They are not meant to be literary works in their own right.
For example, if a user writes, “Summarize this article in formal academic language with three headings,” the prompt is clearly functional. It tells the AI what to do. It does not seek to communicate a creative literary message. In such a case, the prompt resembles a method of control rather than an expressive work. Indian copyright law is unlikely to protect such language because doing so would confer monopoly over ordinary instructions, which are meant to remain freely usable.
This distinction also protects the public domain. If prompts were broadly protected, common language structures, standard commands, and reusable prompt patterns could become subject to private control. That would be inconsistent with the fundamental structure of copyright law. The purpose of copyright is to encourage creativity, not to prevent others from using ordinary functional phrases. Therefore, from the Indian perspective, the more a prompt looks like an instruction, the less likely it is to be protected.
Human Authorship and AI Assistance
A further limitation concerns authorship. Copyright law in India, as elsewhere, is fundamentally human-centered. Although the Act recognizes computer-generated works, it does so only by assigning authorship to the person who causes the work to be created. This raises the question whether the user who writes the prompt can be treated as the author of the prompt and of the resulting output. The answer must be separated into two parts.
First, the prompt itself may be authored by a human. That is not the problem. The issue is whether the prompt contains protectable expression. If it does not, authorship alone cannot rescue it. Second, the output generated by the AI may or may not be copyrightable depending on how much human input shaped it. If the user merely provided a basic prompt and the machine independently created the content, the claim to authorship becomes weak. If the user edited, selected, curated, and substantially transformed the result, the case becomes stronger.
This distinction is important because many commentators mistakenly assume that writing a prompt automatically makes the user the owner of whatever the AI produces. Indian law does not support such a broad claim. The statutory language in Section 2(d)(vi) helps only where the work is truly computer-generated and causation can be tied to a human creator. It does not create copyright in every input string. Thus, the authorship issue is another area where protection is limited.
One might argue that prompts are clearly fixed because they are written down, typed into a platform, saved in a document, or stored in digital form. That is true, but fixation alone is not enough. Copyright law requires both fixation and originality. A prompt can be fixed and still remain uncopyrightable if it lacks the necessary expressive content.
This is a subtle but important point. Many legal disputes begin with the assumption that because something is written, it must be protected. That is not the law. A grocery list, a flight instruction, a common phrase, or a short functional message is fixed but not necessarily original. Prompts often belong to this category. Even long and technically sophisticated prompts may still be viewed as practical instructions unless they reflect genuine literary composition.
The limitation becomes even clearer when the prompt is part of a workflow rather than a standalone text. For instance, prompts used in coding, research summarization, legal drafting, or image generation often operate as internal tools. Their value lies in the result they help produce. Copyright law, however, is concerned with the textual form itself. If the text has no independent expressive quality, fixation does not change the result.
Indian Judicial Approach and Persuasive Case Law
Although Indian courts have not yet delivered a definitive ruling on the copyrightability of AI prompts, existing case law strongly suggests a restrictive approach. The most relevant authority remains Eastern Book Company v. D.B. Modak, where the Supreme Court insisted on creativity beyond effort. That case has become the cornerstone for evaluating originality in India and would likely be applied to any dispute involving prompt engineering.
A second important principle comes from the long-standing rule that copyright does not protect ideas, methods, or systems. Indian courts have repeatedly recognized that copyright is not a monopoly over useful methods of operation. Prompts, which are often operational commands to a machine, fit uncomfortably within this framework. If courts treat them as functional language, they will likely deny protection.
There is also increasing academic and comparative discussion around AI-generated works and prompt-based originality, including debates on whether especially creative prompts could receive a narrow form of protection. But even these discussions do not support a broad rule. At most, they suggest that a highly individualized prompt with literary style, narrative framing, or expressive arrangement might cross the originality threshold. That would be the exception, not the norm.
Policy Concerns and Practical Risks
Beyond doctrine, there are important policy reasons for limiting copyright protection in this area. If prompts were broadly protected, individuals and companies might try to claim ownership over ordinary prompt templates, common command structures, and widely used formulas. This could create “prompt trolling,” where parties use copyright strategically to block others from using similar or identical instructions. Such a development would be harmful to innovation.
There is also a concern that overprotection would discourage experimentation. AI use thrives on iteration, remixing, and shared practices. Researchers, students, content creators, and software developers routinely modify each other’s prompts to improve results. If those materials became heavily protected, the free flow of practical knowledge could be restricted. That would run against the collaborative and experimental nature of AI development.
Another practical concern is evidentiary. Even if a prompt were potentially original, proving authorship and creativity would not be easy. Users may create and discard hundreds of prompts before arriving at a usable one. The process is often iterative and informal. In such conditions, it is difficult to identify the exact original expression that should be protected. This further weakens the case for broad copyright.
From the Indian legal perspective, the copyrightability of AI prompts is highly limited. The Copyright Act, 1957 provides the basic framework through Sections 2(o), 2(d), 13, 17, and related provisions, but the decisive doctrine is originality as interpreted in Eastern Book Company v. D.B. Modak. Most prompts are functional instructions rather than original literary works. They are created to guide a machine, not to stand as independent expressive compositions.
The idea-expression distinction, the human authorship requirement, and the policy against monopolizing ordinary language all point in the same direction. A prompt may be written, stored, and used creatively, but that alone does not make it copyrightable. Only a narrow class of prompts, if any, that display distinctive literary expression and independent creativity might qualify. For all practical purposes, the safer conclusion is that AI prompts remain mostly outside copyright protection in India.
This does not mean prompt engineering lacks value. On the contrary, prompts are increasingly important tools in research, writing, law, design, and business. But their value is primarily functional, strategic, and technical rather than proprietary in the copyright sense. Indian law is likely to protect the expression of a prompt only in rare cases where the language itself rises to the level of original literary authorship. Until the legislature or courts provide more direct guidance, that narrow interpretation remains the most defensible one.