1. Introduction

AI copyright attribution claims concern disputes about whether a person who creates, edits, commissions, or publishes AI-assisted content is entitled to be identified as the author, whether an AI system can itself receive copyright attribution, whether attribution can be removed or falsely assigned, and whether AI-generated material infringes another person's copyright or moral rights.

Indian copyright law does not presently contain a standalone statutory cause of action called an “AI copyright attribution claim.” Such disputes are addressed through the existing framework of the Copyright Act, 1957, particularly:

• authorship and ownership;
• moral rights;
• infringement;
• false attribution;
• substantial reproduction;
• commissioned works;
• employment;
• assignments and licences;
• technological protection measures;
• contractual obligations; and
• constitutional and equitable remedies where appropriate.

The central difficulty is that Indian copyright law is fundamentally human-authorship oriented, while generative AI may produce text, images, music, software, or other works with varying degrees of human involvement.

2. What Is an AI Copyright Attribution Claim?

An AI copyright attribution claim can arise where there is a dispute concerning who should be identified as the creator or author of an AI-assisted work.

Typical disputes include:

1. A human claims authorship of an AI-assisted image.
2. An AI-generated work is published under the name of a person who made little or no creative contribution.
3. A creator's name is removed from an AI-assisted adaptation.
4. AI substantially modifies an existing copyrighted work and the original author alleges improper attribution.
5. A company claims copyright in an AI-generated work created by an employee.
6. A person falsely represents themselves as the author of another person's AI-assisted work.
7. An AI-generated work imitates a living artist's style and is commercially released under a misleading attribution.
8. A publisher or platform removes an author's attribution from an AI-edited work.
9. AI-generated material is incorporated into a human-authored work and the parties dispute who owns the resulting copyright.
10. A person alleges that an AI system has reproduced their copyrighted work and improperly attributed it to someone else.

Thus, “attribution” and “copyright ownership” are related but not identical questions.

3. The Human-Authorship Problem

The most important legal issue is:

Can an AI system itself be the “author” of a copyright work under Indian law?

The practical answer under the present Indian statutory framework is highly doubtful.

The Copyright Act, 1957 repeatedly operates through concepts such as:

• author;
• owner;
• assignment;
• employment;
• licence;
• infringement; and
• moral rights.

These concepts presuppose legal persons capable of holding rights and obligations.

An AI model is not presently recognised in Indian copyright law as an independent legal author capable of owning copyright.

Therefore, the more important question becomes:

Which human being made sufficient creative contribution to the resulting work?

4. AI-Assisted Work vs Purely AI-Generated Work

This distinction is fundamental.

A. AI-assisted human creation

A human may:

• develop the concept;
• select the subject;
• formulate detailed prompts;
• generate multiple outputs;
• select particular outputs;
• edit the output;
• combine several generations;
• make compositional decisions;
• retouch the work;
• arrange elements;
• rewrite AI-generated text;
• add original material; and
• determine the final expressive form.

The stronger the human creative contribution, the stronger the argument for copyright protection and attribution.

B. Purely autonomous AI output

Suppose a user merely enters:

“Generate a picture of a mountain at sunset.”

The AI independently produces the complete image and the user performs no meaningful creative selection or modification.

The argument that the user is automatically the copyright author is substantially weaker.

The mere operation of software should not automatically be equated with authorship.

5. Copyright Act, 1957 — Relevant Framework

Several provisions become particularly important.

Section 2(d) — “Author”

Section 2(d) identifies the author of different categories of works.

For example, in relation to:

• literary works;
• dramatic works;
• musical works;
• artistic works;
• cinematograph films; and
• sound recordings,

the Act specifies who is treated as the author.

The provision is important because an AI attribution dispute ultimately requires identification of the legally recognised author.

Section 13 — Works in Which Copyright Subsists

Section 13 identifies categories of works in which copyright subsists.

Copyright protection therefore depends upon the work satisfying statutory requirements.

An AI output does not automatically receive copyright merely because it was generated by sophisticated technology.

6. Section 17 — First Ownership

Section 17 generally establishes the author as the first owner of copyright, subject to statutory exceptions.

This creates an important question for AI-assisted works:

If the AI is not a legal author, who is the human author?

Potential candidates may include:

• the person who generated the work;
• the person who creatively selected and modified the output;
• an employee;
• an employer;
• a commissioning party; or
• another person determined by contract and the statutory rules.

The answer depends heavily on the facts.

7. Moral Rights and Attribution

One of the most important provisions is Section 57 of the Copyright Act.

Section 57 protects the author's special rights, commonly described as moral rights.

These include the right to claim authorship and protection against certain forms of distortion or mutilation prejudicial to the author's honour or reputation.

This makes Section 57 especially significant for AI attribution disputes.

Example

An artist creates an original illustration and uses AI only for background generation.

The artist then:

• selects the composition;
• redraws the characters;
• changes lighting;
• makes extensive modifications; and
• creates the final work.

If a platform publishes the final work while deliberately removing the artist's name and falsely attributing it to another person, a Section 57 issue may arise.

8. False Attribution

AI technology creates unusual possibilities for false attribution.

For example:

A person generates an AI image in the style of an artist and publishes it under the artist's actual name, creating the impression that the artist created the work.

Potential legal consequences may involve:

• copyright law;
• moral rights;
• passing off;
• consumer protection;
• trademark law;
• defamation;
• contractual liability; and
• potentially other statutory offences depending on the conduct.

The key issue is not simply that AI was used.

The legal problem is the false representation concerning authorship or origin.

9. AI and Existing Copyright Works

Another major category involves AI-generated works that reproduce or closely resemble pre-existing copyrighted material.

Consider:

A user uploads a copyrighted illustration into an AI system and instructs it to create a substantially similar image.

If the resulting work reproduces protected expression, questions may arise concerning:

• reproduction;
• adaptation;
• communication to the public;
• publication;
• substantial similarity;
• fair dealing;
• licence;
• technological reproduction; and
• ownership of the resulting output.

The existence of AI does not eliminate ordinary copyright principles.

10. AI Training and Copyright Attribution

A separate but related problem is AI training.

Suppose an AI company collects millions of copyrighted books, photographs, paintings, or articles to train a model.

Potential disputes include:

• whether copying during training constitutes infringement;
• whether temporary or intermediate copies are legally relevant;
• whether the use falls within a statutory exception;
• whether licences were obtained;
• whether outputs reproduce protected expression;
• whether creators receive attribution; and
• whether creators have contractual or statutory rights concerning the use.

However, training disputes and attribution disputes should not be conflated.

A copyright owner may object to training even if the eventual AI output does not reproduce the original work.

Conversely, an attribution claim may arise even where the primary dispute is about the identity of the author of the output.

11. AI Copyright Attribution and Moral Rights

Moral rights are particularly important because they can survive independently of ordinary economic ownership.

An artist might assign economic copyright rights to a company while retaining statutory moral-rights interests.

For AI-assisted works, questions may arise concerning:

Right of attribution

Who should be identified as the author?

Right against distortion

Has AI alteration changed the work in a manner prejudicial to the author's honour or reputation?

Integrity of the work

Was an original work manipulated through AI so substantially that the creator objects?

Misrepresentation

Has AI-generated material been presented as though it were the original artist's own creation?

12. Case Law

Indian courts have not yet developed a comprehensive Supreme Court jurisprudence specifically deciding generative-AI authorship and attribution.

Accordingly, existing copyright and moral-rights decisions provide the principles that would likely govern an AI dispute.

1. R.G. Anand v. Deluxe Films

(1978) 4 SCC 118

This is one of India's leading copyright cases concerning substantial similarity.

The Supreme Court explained the distinction between an idea and its expression and examined whether the defendant had copied the protected expression of the plaintiff's work.

Relevance to AI

If an AI-generated work allegedly copies an existing copyrighted work, the question remains whether protected expression has been reproduced.

AI does not create a special exemption from the ordinary idea-expression distinction.

Principle: Similarity of ideas alone is insufficient; unlawful copying of protected expression is the critical issue.

2. Eastern Book Company v. D.B. Modak

(2008) 1 SCC 1

The Supreme Court examined originality in copyright law.

The Court rejected an excessively low threshold and emphasised the importance of intellectual effort and originality.

Relevance to AI

This case is particularly important for AI-assisted works.

Where a human merely presses a button and accepts an automatically generated output, the originality attributable to that human may be questioned.

But where the human exercises substantial creative judgment, selection, arrangement, editing and modification, the originality requirement may be satisfied.

Principle: Copyright requires the relevant degree of originality; mechanical labour alone is not enough.

3. Amar Nath Sehgal v. Union of India

2005 (30) PTC 253 (Del)

The Delhi High Court delivered a landmark decision concerning moral rights under Section 57.

The case concerned destruction and treatment of a significant artistic work and recognised the importance of the author's special rights.

Relevance to AI

This is highly relevant where AI:

• substantially alters an artist's work;
• removes the creator's identity;
• modifies an artistic work in a prejudicial manner; or
• creates a distorted version that affects the author's reputation.

Principle: Copyright law protects interests extending beyond purely economic ownership.

4. Mannu Bhandari v. Kala Vikas Pictures Pvt. Ltd.

1987 (1) SCC 503

The Supreme Court considered the author's rights concerning adaptation of a literary work into another medium.

The case illustrates the importance of the author's interests when a work is adapted or transformed.

Relevance to AI

AI can produce:

• adaptations;
• translations;
• transformations;
• illustrations;
• screenplays; and
• derivative content.

Where a copyrighted work is transformed through AI, the underlying author's rights cannot simply be ignored because a machine performed the transformation.

Principle: Adaptation does not automatically extinguish the rights of the original author.

5. Indian Performing Right Society Ltd. v. Eastern India Motion Pictures Association

(1977) 2 SCC 820

The Supreme Court examined copyright interests in musical and literary works incorporated into cinematographic films.

The decision demonstrates that different creative contributions can generate distinct copyright interests.

Relevance to AI

An AI-assisted production may contain contributions from:

• a writer;
• composer;
• photographer;
• illustrator;
• AI operator;
• editor;
• filmmaker; and
• software developer.

The fact that AI participates in production does not automatically eliminate the copyright interests of human contributors.

Principle: Copyright can attach to distinct protected contributions even when those contributions are incorporated into a larger work.

6. Super Cassettes Industries Ltd. v. Bathla Cassette Industries Pvt. Ltd.

(2003) 27 PTC 231 (Delhi)

The Delhi High Court addressed copyright infringement and the importance of examining the nature and extent of copying.

Relevance to AI

AI-generated material can involve extensive copying from source materials.

The court would still need to determine:

• what material was copied;
• whether it was protected;
• how substantial the copying was;
• whether there was a licence; and
• whether an exception applies.

Principle: Copyright infringement depends upon the legally protected expression and the nature and extent of copying.

7. S. Rangarajan v. P. Jagjivan Ram

(1989) 2 SCC 574

Although principally concerning freedom of expression, the case is useful when AI-generated content creates disputes involving artistic expression and public dissemination.

The Supreme Court emphasised the constitutional protection of expression.

Relevance to AI

AI-generated creative works can raise conflicts between:

• copyright;
• artistic freedom;
• speech;
• reputation; and
• public dissemination.

The presence of AI does not remove constitutional protections for lawful expression.

8. Shreya Singhal v. Union of India

(2015) 5 SCC 1

The Supreme Court addressed online speech and intermediary liability.

Relevance to AI

AI copyright disputes frequently occur on:

• social media;
• AI platforms;
• image-sharing websites;
• content platforms; and
• online marketplaces.

The decision provides important background concerning the legal treatment of online intermediaries.

It does not, however, establish that platforms are automatically liable for every AI-generated copyright violation.

13. What These Cases Establish Collectively

The above authorities support several propositions relevant to AI.

14. Who Can Claim Attribution in an AI-Assisted Work?

Several possibilities must be distinguished.

Scenario 1 — Human creates and AI merely assists

The human has the strongest attribution claim.

Scenario 2 — Human heavily edits AI output

The human may claim copyright in the human-created elements and potentially in the resulting compilation/adaptation, depending on originality.

Scenario 3 — Human merely enters a short prompt

The claim is considerably less certain.

The prompt itself may potentially be protected if sufficiently original, but that does not automatically mean the entire AI-generated output belongs to the prompt writer.

Scenario 4 — Employer controls AI production

Employment and contractual principles may affect ownership.

Scenario 5 — Commissioned AI work

The contract becomes extremely important.

Scenario 6 — AI output incorporates pre-existing copyrighted material

The original copyright owner's rights remain relevant.

15. Attribution vs Ownership

These concepts should never be treated as identical.

Attribution

Answers:

Who should be identified as the author?

Ownership

Answers:

Who owns the economic copyright?

Moral rights

Answer:

What continuing personal rights does the author possess concerning authorship and integrity?

For example:

A company might own economic copyright under a valid employment or assignment arrangement, while the human creator may retain moral-rights interests.

Therefore:

Ownership of copyright does not necessarily resolve every attribution question.

16. AI Prompting and Copyright

An increasingly important issue is whether a prompt constitutes a copyrightable creative contribution.

A simple prompt such as:

“A red flower in a garden.”

would ordinarily present a weak originality argument.

But a highly elaborate creative instruction containing:

• detailed narrative;
• character development;
• composition;
• artistic direction;
• dialogue;
• sequencing;
• symbolism;
• colour relationships;
• camera positioning; and
• detailed stylistic instructions

may involve substantially greater intellectual contribution.

However, even then, the legal question is not:

“Was the prompt long?”

The real question is:

Did the claimant contribute sufficient original expression to the protectable final work?

Length alone does not establish originality.

17. Selection and Arrangement

A person may obtain stronger copyright arguments through creative selection and arrangement.

For example, an AI system generates 500 images.

The user:

1. selects 5;
2. combines them;
3. crops particular elements;
4. alters lighting;
5. digitally paints portions;
6. adds original characters;
7. rearranges the composition; and
8. creates a final poster.

The final work may contain protectable human-created expression even if AI generated some underlying material.

This creates a potentially important distinction between:

AI-generated material and human-authored contributions incorporated into the final work.

18. Copyright in AI-Generated Text

The same issue applies to text.

Suppose a person asks AI:

“Write a 2,000-word article about Indian constitutional law.”

If the user publishes the output without meaningful modification, the claim of human authorship over the AI-generated prose is less straightforward.

But if the person:

• develops the argument;
• researches authorities;
• restructures the article;
• rewrites passages;
• adds original analysis;
• selects quotations;
• creates the organisation; and
• substantially edits the language,

the resulting work may contain significant human-authored material.

19. AI-Generated Music

Music creates additional complexity because separate rights may exist in:

• lyrics;
• musical composition;
• sound recording;
• performance;
• arrangement.

If AI generates the melody but a human writes the lyrics and substantially arranges the music, attribution must be analysed component by component.

One person need not necessarily be the legal author of every copyrightable component.

20. AI-Generated Software

AI-assisted programming raises questions concerning:

• source code;
• algorithms;
• documentation;
• architecture;
• databases;
• open-source components;
• third-party code; and
• employment ownership.

A developer who uses AI to generate portions of code may still make significant human contributions through:

• architecture;
• debugging;
• selection;
• integration;
• modification;
• testing; and
• implementation.

But copying AI-generated code from another protected program can create infringement concerns.

21. AI and False Attribution of an Existing Artist

Consider:

A company produces an AI-generated painting and labels it “by M.F. Husain.”

That raises a very different issue from ordinary AI authorship.

The company is not merely asking:

“Who owns this image?”

It is making a representation concerning the identity of the creator.

Depending on the circumstances, possible legal theories could include:

• copyright/moral-rights principles;
• passing off;
• personality/publicity interests where applicable;
• trademark law;
• consumer protection;
• defamation; and
• contractual or commercial liability.

22. AI Style Imitation

Another difficult issue is:

Is an AI-generated work unlawful merely because it imitates an artist's style?

Generally, style as such is not equivalent to copyright-protected expression.

Copyright protects expression rather than an abstract artistic style.

Therefore:

“Create a painting inspired by impressionist techniques”

is legally different from:

“Reproduce this specific copyrighted painting with only minor changes.”

The factual degree of copying matters.

23. Right of Attribution and Section 57

Section 57 deserves special attention in AI disputes.

A creator may potentially object where:

• their name is removed;
• another person is falsely identified as the creator;
• their work is distorted;
• their work is mutilated;
• an alteration prejudicially affects their honour or reputation.

The AI context may make the factual question more complicated because the modification may have been performed automatically.

But automation does not necessarily defeat the underlying legal right.

24. Liability of the AI Developer

The AI developer is not automatically liable for every copyright attribution dispute involving an output.

Liability may depend upon:

• what the developer represented;
• whether the system was designed to reproduce protected material;
• terms of service;
• licences;
• notice-and-takedown mechanisms;
• control over outputs;
• knowledge;
• contractual commitments;
• statutory duties; and
• the developer's actual role in the alleged infringement.

The developer, platform, user and commercial publisher may therefore have different legal positions.

25. Liability of the AI User

The user may face liability where the user:

• deliberately uploads copyrighted material;
• instructs the system to reproduce it;
• commercially publishes infringing material;
• removes attribution;
• falsely claims authorship;
• knowingly distributes infringing material; or
• breaches a licence.

The fact that “AI generated it” is not necessarily a complete defence.

26. Liability of an Employer

Where an employee uses AI during employment, ownership may depend upon:

• Section 17;
• employment relationship;
• job duties;
• contractual provisions;
• company AI policies;
• the employee's creative contribution; and
• the nature of the work.

A company cannot simply assume that every AI output created by an employee automatically belongs to it without examining the applicable statutory and contractual framework.

27. Commissioned AI Works

Contracts should ideally specify:

1. who controls the AI tool;
2. who owns the output;
3. who is responsible for copyright clearance;
4. who warrants originality;
5. who provides source materials;
6. whether AI use must be disclosed;
7. whether human authorship is required;
8. attribution requirements;
9. moral-rights obligations;
10. indemnities;
11. platform restrictions;
12. training-data restrictions; and
13. remedies for infringement.

Without contractual clarity, attribution disputes can become extremely difficult.

28. Evidence in AI Attribution Litigation

A claimant should preserve:

AI interaction records

• prompts;
• outputs;
• revisions;
• timestamps;
• conversation history.

Human creative records

• sketches;
• drafts;
• source files;
• editing history;
• Photoshop/Illustrator files;
• manuscripts;
• development notes.

Platform records

• account information;
• publication history;
• metadata;
• version history.

Contractual records

• employment agreement;
• commissioning contract;
• licence;
• AI platform terms;
• assignment agreements.

Technical evidence

• model/version;
• generation parameters;
• image metadata;
• provenance records;
• content credentials where available.

These can become crucial in proving who actually contributed creatively.

29. Burden of Proof

A claimant alleging infringement generally needs to establish the relevant copyright interest and the infringing conduct.

In an attribution dispute, evidence may additionally need to establish:

1. existence of the claimant's work;
2. claimant's authorship;
3. originality;
4. human creative contribution;
5. ownership or relevant rights;
6. defendant's conduct;
7. similarity or copying where relevant;
8. removal or falsification of attribution;
9. prejudice where required for moral-rights relief; and
10. resulting damage or entitlement to relief.

30. Defences

Possible defences include:

No copyright subsists

The alleged work may lack the required originality.

No human authorship

The claimant may not establish sufficient human creative contribution.

Independent creation

The defendant independently created the material.

No substantial copying

Similarity may concern ideas, concepts or unprotected elements.

Fair dealing

The use may fall within a statutory exception where applicable.

Licence

The defendant may have permission.

No attribution obligation

The contractual arrangement may not require the alleged attribution.

No prejudice

A moral-rights claim may fail if the statutory requirements concerning prejudice are not established.

Intermediary position

A platform may argue that its role and legal obligations differ from those of the person who generated or uploaded the material.

31. Remedies

Depending upon the claim, remedies can include:

Injunction

The court may restrain continuing infringement or wrongful use.

Damages

Compensation may be available for established loss.

Account of profits

In appropriate cases, the claimant may seek the defendant's profits attributable to the wrongful conduct.

Delivery up/destruction

Infringing copies and related materials may be subject to appropriate orders.

Declaration

A court may determine authorship or copyright entitlement.

Attribution/correction

Where legally appropriate, relief may address false attribution or removal of authorship.

Moral-rights relief

Section 57 can be particularly important where distortion or false attribution affects the author's interests.

32. AI Attribution — Practical Legal Test

A useful framework for Indian litigation is:

Step 1 — Identify the work

Is it:

• literary;
• artistic;
• musical;
• dramatic;
• software;
• cinematographic;
• sound recording; or another protected category?

Step 2 — Identify the human contribution

Who actually made creative decisions?

Step 3 — Determine originality

Is there sufficient original expression?

Step 4 — Determine ownership

Who owns the relevant copyright under Section 17, assignment, employment or contract?

Step 5 — Examine AI involvement

What exactly did the AI generate?

Step 6 — Separate AI material from human material

Which elements were machine-generated and which were created or substantially modified by humans?

Step 7 — Examine attribution

Was the correct author identified?

Step 8 — Examine alteration

Was the original work modified or distorted?

Step 9 — Examine copying

Does the AI output reproduce protected expression belonging to someone else?

Step 10 — Determine remedy

Consider injunction, damages, account of profits, declaration, attribution-related relief and other statutory remedies.

33. Important Distinction: AI Output Does Not Automatically Belong to the User

One of the most common misconceptions is:

“I typed the prompt, therefore I automatically own the copyright.”

That proposition is not established Indian law.

The user's legal position depends on:

• the nature of the human contribution;
• originality;
• the resulting work;
• applicable contractual terms;
• the statutory framework; and
• the rights of other contributors.

Likewise:

“AI generated it, therefore nobody can own copyright.”

That is also too broad.

A final work may contain substantial protectable human expression even where AI was heavily involved.

34. Six Core Case Laws — Quick Reference

Additional useful authorities include S. Rangarajan v. P. Jagjivan Ram, (1989) 2 SCC 574 and Shreya Singhal v. Union of India, (2015) 5 SCC 1.

35. Emerging Indian Legal Position

As of 2026, Indian courts have not yet produced a definitive Supreme Court judgment establishing a comprehensive test for:

“Who is the copyright author of a purely generative-AI output?”

Consequently, courts are likely to apply existing statutory concepts rather than create an entirely separate category of AI copyright.

The most important principles are likely to be:

1. AI is a tool, not presently a recognised copyright author.
2. Human creative contribution remains central.
3. A prompt alone does not automatically confer copyright over the resulting output.
4. Human selection, arrangement, editing and transformation may create protectable rights.
5. Existing copyright in source material remains relevant.
6. AI does not eliminate moral rights.
7. False attribution can create liability independent of ordinary economic ownership.
8. Contracts and employment arrangements can materially change ownership questions.
9. AI developers, users, employers and publishers may have different liabilities.
10. The ultimate question is fact-specific and depends on the nature of the human contribution and the alleged infringement.

36. Conclusion

AI Copyright Attribution Claims in India are presently best understood as an application of existing copyright law to AI-assisted creation rather than as a separate statutory cause of action.

The most significant legal questions are:

Who made the legally relevant creative contribution?

Is the resulting work sufficiently original?

Who owns the copyright?

Who is entitled to attribution?

Has another person's protected expression been copied?

Has the work been distorted or falsely attributed?

The strongest legal foundation for attribution disputes is therefore found in Sections 2(d), 13, 17 and 57 of the Copyright Act, 1957, supplemented by the law governing infringement, assignments, employment, contracts, moral rights and remedies.

The cases of R.G. Anand, Eastern Book Company, Amar Nath Sehgal, Mannu Bhandari, IPRS and Super Cassettes provide the principal doctrinal building blocks, although none was decided specifically on modern generative AI. Consequently, any claim involving AI should carefully distinguish AI-generated material, human-authored material, pre-existing copyrighted material, ownership, and moral-rights attribution rather than treating all AI output as a single legal category.