CLAT 2027 cohort — enrolment closes soonJudiciary Foundation now enrollingSummer litigation internship — apply by 5 JulyCorporate & transactional internship — 8 seatsLegal Drafting Intensive — 4 weeks, small cohort
Izel.Legal
← All articles
Practice8 min read

Protection of AI-Generated Works under Indian Copyright Law: Rethinking the Requirement of Human Authorship

The central issue, therefore, is not simply whether AI was used in producing a work. The more important question is whether there was sufficient human creativity and control behind the final work to justify granting copyright protection.

PS
Prachi Shrivastava
Research & Content Team ·

Introduction

Artificial Intelligence (AI) has changed the way creative works are produced. Today, a person can give a simple instruction to an AI system and, within seconds, receive a poem, photograph, painting, song, video or written article. This development has created a difficult question for copyright law: when a machine produces a creative work, who should be considered its author?

Traditionally, copyright law has been built around the idea of human creativity. A writer writes a book, a musician composes a song, a painter creates a painting and a photographer takes a photograph. The law gives these individuals certain exclusive rights because they have contributed their creativity to the work. AI challenges this traditional understanding because a person may now create a work without personally producing every element of its final expression.

For example, a person may simply type, “Create a realistic painting of an Indian village during monsoon,” and an AI system may independently decide the colours, people, background, lighting and overall composition. In such a situation, it becomes difficult to say how much of the final creative expression actually comes from the human user.

Indian copyright law does not contain a detailed statutory framework specifically dealing with generative AI. However, the Copyright Act, 1957 does contain an interesting provision concerning computer-generated works. Section 2(d)(vi) states that, in relation to a literary, dramatic, musical or artistic work which is computer-generated, the author is “the person who causes the work to be created.”1 This provision provides a starting point for analysing AI-generated works, but it was introduced long before the emergence of today's sophisticated generative AI systems.

The central issue, therefore, is not simply whether AI was used in producing a work. The more important question is whether there was sufficient human creativity and control behind the final work to justify granting copyright protection.

Understanding AI-Generated Works

AI-generated works are creative outputs produced wholly or partly through an Artificial Intelligence system. They may include literary works, paintings, photographs, music, videos, designs and other forms of creative expression.

However, there is an important difference between AI-assisted works and AI-generated works.
In an AI-assisted work, the human remains the main creator and uses AI merely as a tool. For instance, a photographer may take a photograph himself and then use an AI tool to remove unwanted objects or improve the lighting. Similarly, a writer may write an article himself and use AI only to correct grammar.

In contrast, in a heavily AI-generated work, the person may provide only a short prompt and allow the AI system to determine most of the creative elements.This distinction matters because copyright protects the expression of human creativity, and not merely the fact that a person operated a computer.

The Supreme Court of India considered the meaning of originality in Eastern Book Company v DB Modak. The Court made it clear that copyright does not arise merely because someone has spent considerable time, labour or money on a work. The work must contain at least a minimum degree of creativity and must originate from the author.2

This principle is particularly relevant to AI-generated works. A person who merely presses a button after entering a basic instruction may have a weaker claim than a person who exercises substantial creative control over the final work.

Copyright Protection under Indian Law

Section 13 of the Copyright Act, 1957 provides copyright protection to original literary, dramatic, musical and artistic works, as well as cinematograph films and sound recordings.3

The requirement of originality is therefore important. However, originality in copyright law does not mean that the work must be completely new or revolutionary. In Eastern Book Company, the Supreme Court explained that copyright law does not demand novelty or invention in the same manner as patent law. What is required is some meaningful creative contribution.4


The Court specifically observed that there must be a minimal degree of creativity and that a purely mechanical exercise is not sufficient.5 This creates an interesting situation for AI.

Suppose a user provides a detailed description, repeatedly modifies the instructions, chooses between numerous outputs and then substantially edits the selected output. There is a stronger argument that the final work reflects the person's own creative choices.On the other hand, where the user merely enters a very general instruction and accepts the first output generated by the AI, it becomes more difficult to identify sufficient human creativity.
Therefore, the mere use of AI should not automatically destroy copyright protection. What should matter is the nature and extent of human contribution.

The Meaning of “Author” under Indian Copyright Law

The most important statutory provision for AI-generated works is Section 2(d) of the Copyright Act.
Section 2(d)(vi) specifically provides that, in relation to a literary, dramatic, musical or artistic work which is computer-generated, the author is the person who causes the work to be created.6
This provision is significant because Indian law already recognises the possibility of a work being generated through a computer.

However, the provision creates another question: who is the person who “causes” the work to be created when generative AI is involved?

Consider two examples.

  • In the first example, an artist gives detailed instructions to an AI system, generates several versions, selects specific elements, repeatedly modifies the prompts and then manually edits the final image.

  • In the second example, a person enters the words “make a beautiful sunset painting” and simply downloads whatever the AI produces.

In both situations, the person technically caused the AI system to produce the work. However, the level of creative involvement is clearly different. A broad interpretation of Section 2(d)(vi) could mean that simply initiating the AI process is enough to make someone an author. But such an interpretation could conflict with the originality requirement recognised by the Supreme Court. Therefore, Section 2(d)(vi) should ideally be read together with Section 13 and the judicial understanding of originality.

Does Indian Copyright Law Require Human Authorship?

The Copyright Act does not expressly use the phrase “human authorship” as a separate statutory requirement. However, the structure of copyright law strongly connects authorship with human intellectual and creative contribution.

At the same time, Section 2(d)(vi) specifically refers to the person who causes a computer-generated work to be created.7 This suggests that even where a computer is involved in generating the work, Indian law contemplates a human author rather than recognising the machine itself as the author. This is important because an AI system does not have legal personality in the same way as a human being or a recognised legal entity. It cannot independently exercise copyright, enter into an assignment, enforce moral rights or sue for infringement in its own name. Therefore, treating AI itself as the copyright owner would require a significant change in the existing legal framework. The more practical approach is to identify the human being who made a sufficiently meaningful creative contribution to the final work.

The RAGHAV AI Controversy

An important development in India concerning AI and copyright involved RAGHAV, an Artificial Intelligence-powered painting application. Applications were made before the Indian Copyright Office concerning an artwork generated using RAGHAV. In one application, the AI system was identified as the author. The application was subsequently questioned by the Copyright Office, demonstrating the difficulty of treating an AI system as an “author” under the existing copyright framework.

A later application involved a human being along with RAGHAV, raising the question of whether AI could be treated as a co-author. The controversy highlighted the uncertainty surrounding the application of Section 2(d)(vi) to modern AI systems.8 The RAGHAV episode is particularly important because it demonstrates that Indian copyright law has not yet developed a clear and consistent approach to AI-generated works.
It also shows why simply declaring an AI system to be an “author” may not solve the problem. The existing law is based upon the assumption that the author is capable of holding and exercising legal rights.

The Importance of Human Creativity

The Supreme Court's judgment in Eastern Book Company v DB Modak provides a useful legal test for analysing AI-generated works.

The Court rejected the traditional “sweat of the brow” approach under which substantial labour and effort could itself justify copyright protection. Instead, the Court required a minimum level of creativity.9 The significance of this principle becomes clear when AI-generated content is considered. For example, imagine that two people use the same AI system.

Person A enters: “Generate a picture of a woman standing in a garden."
The AI creates the entire image, and Person A uses the first result without making any changes.
Person B, however, develops a detailed artistic concept, experiments with numerous prompts, selects particular visual elements, combines several outputs, changes the composition and finally edits the image manually. Person B has made considerably more creative decisions.

Applying the reasoning in Eastern Book Company, Person B would have a stronger argument that the final work contains an original human contribution. Person A's claim would be considerably weaker because the creative choices may have primarily been made by the AI system.
Thus, the law should focus on human creative control rather than simply human interaction with the AI system.

AI as a Tool or AI as a Creator?

One useful way to understand this issue is to compare AI with traditional creative tools.
A camera is not considered the author of a photograph. A paintbrush is not the author of a painting. Similarly, ordinary editing software does not become the author of a digitally edited photograph.
In these situations, the human makes the creative decisions while the technology assists in executing them.

Generative AI is different because it can itself make numerous creative-looking choices. For example, after receiving a simple prompt, it may decide the appearance of a person, the background, colours, lighting and composition.

This raises the question of whether AI should still be treated merely as a tool. The answer should probably depend on the circumstances.Where AI merely assists the human creator, the human should remain the author. Where AI independently determines almost all expressive elements and the human contribution is minimal, copyright protection for the resulting work becomes more difficult to justify.


Who Should Own Copyright in AI-Generated Works?

If a work qualifies for copyright protection, the next question is ownership. Section 17 of the Copyright Act generally provides that the author of a work is the first owner of copyright, subject to certain statutory exceptions.10 Therefore, if the human user is recognised as the author under Section 2(d)(vi), that person would ordinarily be the first copyright owner, unless another statutory or contractual arrangement applies.

However, AI-generated works may involve several people:

  • The person who gives the prompt

  • The person who edits the output

  • The person who commissions the work

  • The company providing the AI system

  • The developers who created the AI model.

It would not be appropriate to automatically give copyright to the AI developer simply because the developer created the software. For example, the company that develops a word-processing program does not automatically become the copyright owner of every novel written using that program. Similarly, the developer of an AI system should not automatically receive copyright in every creative output generated by users. Ownership should instead be connected to the creative contribution and the legal relationship between the parties.


AI-Generated Works and the Problem of Registration

Copyright protection in India does not generally depend upon registration. Copyright arises automatically when a qualifying work is created. Nevertheless, registration can provide useful evidence concerning ownership and the existence of the work. AI-generated works make this process more complicated.

If a person applies for registration of an AI-generated painting, questions may arise regarding:

  • Who created the work?

  • whether AI was used?

  • How extensively AI was used;what instructions were given?

  • How much the applicant edited the final output?

  • Whether the final work reflects sufficient human creativity?

For this reason, creators who use AI commercially should maintain records of their creative process. Prompts, drafts, different AI outputs, editing stages and the final version can help demonstrate the extent of human involvement. Such documentation could become particularly valuable if ownership is challenged in court.

Copyright and AI Training Data

The copyright problem does not end with the AI-generated output. There is another major question: what material was used to train the AI system?

Generative AI systems may be trained on extremely large datasets containing books, photographs, artworks, music, websites and other materials. Some of these materials may be protected by copyright. This creates a separate legal issue concerning whether the copying or use of copyrighted works during the training process amounts to infringement.

It is important not to confuse two different questions:

First: Is the AI-generated output itself capable of receiving copyright protection?

Second: Did the AI system infringe someone else's copyright while being trained or while generating the output?

These questions require separate legal analysis.

AI and Moral Rights

Indian copyright law also recognises certain special rights of authors under Section 57 of the Copyright Act.11 These are commonly referred to as moral rights and include the right to claim authorship and the right to object to certain prejudicial modifications of the work.
AI creates an interesting problem here. If the AI itself were considered the author, how could it exercise moral rights? A machine cannot meaningfully claim that its reputation has been harmed or that its work has been distorted.
This is another reason why recognising AI itself as an author does not fit comfortably within the existing Indian copyright framework. If a human being is recognised as the author because of substantial creative involvement, however, the existing framework of moral rights can continue to operate more naturally.


Difficulties in the Existing Legal Framework

The biggest problem is that the Copyright Act was not designed with modern generative AI in mind. Section 2(d)(vi) provides a useful starting point, but it does not answer several important questions.

For example:

1. How much human involvement is necessary?

2. Is a detailed prompt sufficient to establish authorship?

3. Who owns an output generated from a jointly written prompt?

4. What happens where AI creates 95% of the final work and the human edits only 5%?

5. Can an AI system be treated as a co-author?

6. How should copyright registration applications disclose AI involvement?

7. Who owns the copyright when an employee creates AI-generated material during employment?

8. How should copyright law deal with AI-generated works that closely resemble existing copyrighted works?

These questions demonstrate that the existing legislation requires clarification.

Rethinking the Requirement of Human Authorship

It would be too simplistic to say that every AI-generated work should be denied copyright protection.

Imagine an artist who spends several weeks developing an idea, writing detailed prompts, generating hundreds of possible outputs, selecting particular elements, rearranging them and manually editing the final work. Although AI has been involved, the final work may still reflect substantial human creativity.
At the same time, automatically granting copyright to every AI output could create another problem. If a person enters a very basic prompt and obtains a sophisticated image created almost entirely by AI, granting that person exclusive rights could give them a monopoly over expression they did not meaningfully create. Therefore, India should adopt a middle approach.

The law should ask: What creative contribution did the human make to the final expression?
The greater the human's creative control, the stronger the claim to copyright should be.
This approach would be consistent with the Supreme Court's emphasis on originality, skill, judgment and minimum creativity in Eastern Book Company.12

Suggested Legal Approach for India

India does not necessarily need to completely rewrite the Copyright Act to address AI. A more practical solution would be to clarify the existing framework.

First, the meaning of “the person who causes the work to be created” under Section 2(d)(vi) should be clarified in the context of generative AI.

Second, the Copyright Office could issue guidelines explaining how AI-assisted works should be treated during registration.

Third, the law could adopt a human creative contribution test. Courts could consider:

  • The nature and detail of the instructions

  • The degree of human control

  • Selection and arrangement of AI outputs

  • Human editing and modification

  • The originality of the final expression

  • Whether the final work reflects independent human creative choices.

Finally, applicants could be encouraged or required to disclose the extent of AI involvement when seeking registration. This would improve transparency without automatically denying copyright protection to AI-assisted works.

Conclusion

Artificial Intelligence has forced copyright law to reconsider one of its most basic assumptions: creative works are normally the product of human beings.

Indian law already provides a useful starting point through Section 2(d)(vi), which identifies the person who causes a computer-generated work to be created as its author. However, the provision was drafted before the emergence of modern generative AI and therefore does not completely answer today's difficult questions.

The most balanced approach would be neither to give copyright directly to AI nor to deny copyright to every work in which AI has been used. Instead, the law should focus on meaningful human creativity.

Where AI is simply being used as a tool and the final work reflects the human user's creative decisions, copyright protection should remain available. On the other hand, where AI independently produces almost the entire expression and the human contribution is limited to a simple instruction, granting full copyright protection becomes much harder to justify.

Ultimately, the real question for Indian copyright law is not simply “Can AI be an author?” The more important question is: “When humans and AI create together, where should the law draw the line between human creativity and machine-generated expression?”

The answer to this question will become increasingly important as AI becomes a regular part of the creative industries. Indian copyright law therefore needs to evolve in a way that protects genuine human creativity, encourages technological innovation and prevents artificial systems from receiving legal rights that the existing copyright framework was never designed to give them.


REFERENCES

[1] The Copyright Act 1957, s 2(d)(vi) (India).

[2] Eastern Book Company v DB Modak (2008) 1 SCC 1. The Supreme Court rejected a purely “sweat of the brow” approach and required a minimum degree of creativity for copyright protection.

[3] The Copyright Act 1957, s 13(1) (India).

[4] Eastern Book Company v DB Modak (2008) 1 SCC 1. The Court clarified that originality does not require novelty or invention; rather, the work must originate from the author and contain the requisite degree of creativity.

[5] Eastern Book Company v DB Modak (2008) 1 SCC 1. The Court held that the exercise of skill and judgment must not be merely trivial or mechanical and that some minimum creativity is required.

[6] The Copyright Act 1957, s 2(d)(vi).

[7] ibid

[8] ibid

[9] Eastern Book Company v DB Modak (2008) 1 SCC 1. The Supreme Court held that merely expending labour, skill and capital is not enough; the work must demonstrate the required degree of creativity.

[10] The Copyright Act 1957, s 17 (India).

[11] The copyright act 1957, s 57 ( india)

[12] Eastern Book Company v DB Modak (2008) 1 SCC 1. The Supreme Court emphasised that copyright requires more than mere labour and that the work must possess sufficient originality and a minimum degree of creativity.

PS
Written by
Prachi Shrivastava
Research & Content Team
Keep reading

More field notes.

Next step

Book a free discovery call.

30 minutes with a senior advocate to scope your matter or your course. We'll tell you honestly whether we can help — and what the right engagement would look like.

Or send us a message →

Next CLAT cohort starts 20 July
Internship applications close 5 July