Who Owns AI-Generated Content? The Copyright Debate Facing Indian Creators

Artificial intelligence has transformed the way creative content is produced. Writers use generative AI to develop articles and marketing copy, designers create illustrations from text prompts, musicians experiment with AI-generated compositions, and filmmakers use automated tools to produce images, voices and video sequences. What once required specialised skills, expensive equipment and considerable time can now be completed in minutes with a few instructions. However, this technological shift has created a difficult legal question: who owns the copyright in content generated by artificial intelligence?

For Indian creators, the question is more than a theoretical debate. It affects freelancers, publishers, advertising agencies, designers, software developers, educators, musicians, production houses and businesses that use AI in their everyday work. If a creator uses an AI tool to produce an image, can they prevent another person from copying it? If a company commissions an AI-generated advertisement, does the company automatically own the rights? Can an AI developer claim ownership simply because its technology produced the result? These questions become even more complicated when AI systems are trained on books, photographs, music, articles and other copyrighted material.

India’s Copyright Act, 1957 contains a provision dealing with computer-generated works, but it was introduced long before modern generative AI became widely available. Recent developments, including the Copyright Office’s 2026 decision involving the DABUS artificial intelligence system, have brought the debate into sharper focus. The central challenge is to determine how existing copyright principles apply when a machine produces the final expression but a human may have supplied the instructions, creative direction, selection and editing.

Understanding Copyright in the Age of Generative AI

Copyright is a form of intellectual property protection that gives creators and other recognised rights holders certain exclusive rights over original works. In India, the Copyright Act, 1957 protects categories such as literary, dramatic, musical and artistic works, as well as cinematograph films and sound recordings. These protections can cover books, articles, photographs, illustrations, paintings, music, films, software and other qualifying creative expressions.

Copyright does not generally protect an idea, theme, fact or general style by itself. It protects the particular expression of an idea. For example, the concept of writing a story about a futuristic city is not owned by one author. However, the original wording, characters, illustrations and other expressive elements of a particular story may qualify for protection. This distinction matters when AI tools produce content based on broad instructions or references to existing creative works.

Traditional copyright law assumes that a person creates or contributes to a work. Generative AI complicates this assumption because the system can produce text, images, music and code by processing patterns learned from large datasets. A user may provide a prompt, but the model determines many details of the final output. The extent of human creative control can therefore vary considerably from one project to another.

An AI-assisted article that a writer extensively researches, restructures, rewrites and edits is different from a complete article generated automatically and published without meaningful human intervention. Similarly, an illustrator who uses AI to explore concepts and then manually develops a final composition has a different creative role from someone who enters a short prompt and accepts the first image produced. Copyright analysis must take these differences into account.

What Does India’s Copyright Act Say About AI-Generated Content?

One of the most important provisions in this debate is Section 2(d)(vi) of the Copyright Act, 1957. It defines the author of a computer-generated literary, dramatic, musical or artistic work as the person who causes the work to be created. This provision is significant because it recognises that a computer-generated work may have an identifiable author even when a computer is involved in its production.

The provision was introduced in an era when computer-generated creative work was very different from the output of today’s generative AI systems. At that time, a person could use computer software to create a design, illustration or other work while exercising substantial control over the process. Modern AI systems, however, can make numerous creative decisions independently after receiving a relatively short instruction. This raises the question of whether the person entering a prompt necessarily qualifies as the person who causes the work to be created.

There are several possible candidates. The user may argue that their instructions, references and creative direction caused the output. A developer might argue that the system’s architecture and training made the output possible. A business that commissioned the work may claim contractual ownership. The answer cannot be determined simply by identifying who paid for the software or who owns the computer on which it was used.

The Act also requires works to satisfy the relevant conditions for copyright protection. Section 13 provides for copyright in original literary, dramatic, musical and artistic works, among other categories. Identifying a possible author under Section 2(d)(vi) does not automatically establish that every machine-generated result is protected. Questions of originality, the nature of human involvement and the particular facts of creation remain important.

The official guidance published by the Intellectual Property Office of India explains that, for a computer-generated work, the author is the person who causes it to be created. However, the practical meaning of this rule for modern generative AI is still developing. The law does not provide a simple formula stating that the person who enters a prompt always owns the resulting content.

Can Artificial Intelligence Be the Legal Author?

A central principle emerging from India’s recent copyright debate is that an AI system cannot simply be named as the legal author in place of a recognised person. Artificial intelligence can generate creative material, but it does not have legal personality in the way an individual or legally recognised entity does. It cannot independently exercise copyright, enter into licensing agreements or enforce ownership rights as a human author or company can.

This issue became particularly visible in the dispute involving US computer scientist Stephen Thaler and his AI system, DABUS. Thaler sought copyright registration in India for an artwork titled A Recent Entrance to Paradise, identifying DABUS as its creator. On 31 August 2026, the Indian Copyright Office rejected the application in which the AI system was named as the author. The decision was reported in September 2026 and clarified that an AI system could not be recognised as the author under the existing framework.

The decision is important because it addresses the distinction between a work’s potential originality and the legal identity of its author. The Copyright Office reportedly accepted that the artwork met the originality requirement but did not accept the AI system as its author. In other words, a work may be considered original without the machine that generated it being legally recognised as the person who owns authorship rights.

The decision does not mean that every work produced with AI is automatically excluded from copyright protection. Rather, it reinforces the need to identify a legally recognised author. It also leaves important questions open, including how much human involvement is required and how Section 2(d)(vi) should apply when a generative AI system performs much of the creative work.

Does the Person Who Writes the Prompt Own the Output?

Prompting is often described as a new creative skill. A user may spend time defining the subject, mood, visual composition, writing style, tone, camera perspective, colour palette and other details. They may also provide reference images, refine instructions repeatedly and select one result from hundreds of alternatives. In some projects, this process can involve considerable creative judgment.

However, the existence of a prompt does not necessarily establish copyright ownership. A short instruction such as “create a photograph of a mountain at sunrise” may leave most expressive decisions to the AI system. The user has specified a general idea, but the system determines the detailed composition, lighting, perspective and other elements. Since copyright generally protects expression rather than ideas, a broad instruction may provide a weak basis for claiming authorship of the entire output.

A more substantial creative process may produce a stronger claim. Consider a designer who develops an original concept, supplies detailed sketches and reference material, generates multiple variations, selects specific elements, combines them with independently created artwork and manually edits the final image. In that situation, the designer’s contribution is more than a simple request for an image. The person has exercised creative control over the resulting expression.

The same distinction applies to writing. A content creator who uses AI to generate an initial draft and then independently develops the argument, verifies facts, rewrites passages, adds original analysis and structures the final article may have copyright interests in the human-authored contributions. The unedited AI-generated portions may require separate consideration. The precise scope of protection will depend on the work and the evidence of human creative contribution.

Therefore, creators should not assume that every prompt creates ownership. The more meaningful and demonstrable their contribution to the final expression, the stronger the basis for explaining their authorship claim.

Who Owns AI-Generated Content: The User, Developer or Company?

The ownership question becomes complicated when several parties participate in the production process. An AI developer creates the technology, a user provides instructions, a business commissions the project and an editor or designer modifies the output. Each party may have a different relationship with the final work.

The AI developer does not automatically own every output generated by its system. Creating a model or providing access to a platform does not necessarily make the developer the author of each individual image, article or musical composition produced by users. However, the platform’s terms of service may establish contractual permissions and restrictions concerning generated content, commercial use, user inputs and platform rights. These terms should be examined separately from copyright ownership under Indian law.

The user may have a stronger claim where they provide meaningful creative direction and make substantial original contributions. Nevertheless, the extent of their control matters. A user who accepts an automatically generated result without meaningful creative intervention may face uncertainty about whether the output qualifies for copyright protection and who should be identified as its author.

Businesses also need to distinguish between commissioning content and owning copyright. If an advertising agency hires a freelancer to create AI-assisted campaign material, the agency should not assume that payment alone transfers every relevant right. The agreement should identify the deliverables, ownership of human-created contributions, permitted use of AI-generated elements, responsibility for third-party material and any restrictions imposed by the tools used.

The Copyright Act contains rules concerning first ownership, employment and commissioned works. Their application depends on the category of work and the circumstances in which it was created. A written contract can clarify the parties’ intentions and allocate rights in protectable contributions, but it cannot automatically turn material that does not qualify for copyright protection into a protected work.

The Copyright Problem Behind AI Training Data

Ownership of AI-generated output is only one part of the debate. Another major issue concerns the material used to train generative AI systems. These systems learn patterns from large collections of text, images, music, software and other data. Some of those materials may be protected by copyright.

Writers, photographers, musicians, publishers and news organisations have raised concerns about their work being collected or used to train AI models without permission or payment. They argue that their creative output helps commercial systems develop capabilities that may compete with the original creators. For example, a photographer may object to images being used to train a model that can generate similar visual material, while a publisher may question whether its articles should be included in a commercial training dataset.

The legal debate involves several distinct activities. Making a copy of a copyrighted work, storing it, processing it for training and generating an output that reproduces protected expression may raise different legal questions. It is therefore important not to treat every use of copyrighted material in AI development as legally identical. The purpose of use, the nature of the material, the relevant statutory exceptions and the facts of the case can all matter.

Section 52 of the Copyright Act provides certain exceptions, including fair dealing for specified purposes. Whether a particular AI training activity falls within an exception is a matter of legal interpretation and cannot be assumed merely because the activity involves technology or research.

In February 2024, the Ministry of Commerce and Industry stated in a parliamentary response that India’s existing intellectual property framework was considered capable of addressing AI-related works and that there was no proposal at that time to create a separate category of AI rights. The same response noted that commercial use of copyrighted works by generative AI users may require permission where the use is not covered by an applicable exception.

The debate has since moved into more detailed policy and litigation discussions. In July 2026, the Delhi High Court issued an interim order in the copyright dispute brought by news agency ANI against OpenAI. The court reportedly refused ANI’s request for an interim injunction and, at the prima facie stage, accepted that the use of the material for model training fell within the fair-dealing exception for research. The dispute concerns the use of copyrighted material for training, not a final determination of who owns AI-generated output. It should not be treated as a comprehensive ruling on every form of AI training or every creator’s rights.

For Indian creators, this distinction is crucial. The question of whether a company can use copyrighted material to train a model is not the same as the question of who owns an image or article generated by that model. Both issues are connected to AI copyright, but they require separate legal analysis.

How AI Copyright Disputes Affect Indian Creators

The uncertainty surrounding AI ownership has practical consequences across India’s creative industries. Independent writers may use AI to research, draft and edit articles, while publishers may incorporate generated content into websites and digital publications. If ownership is disputed, creators may find it difficult to establish exclusive rights over material that contains substantial machine-generated expression.

Graphic designers and illustrators face a related challenge. AI image generators can produce visual material resembling existing artistic styles, characters or compositions. A creator may spend time refining an output only to discover that another user has generated a similar image. Similarity alone does not automatically prove infringement, but the presence of recognisable protected expression can create legal risks.

Musicians and audio producers also face questions about AI-generated melodies, vocals and sound recordings. The use of a person’s voice or a recognisable performance may raise issues beyond copyright, including contractual rights, privacy and personality-related interests. A generated song may involve several different elements, such as lyrics, composition, performance and sound recording, each of which can have separate rights holders.

Software developers must consider the distinction between AI-generated code and code copied from existing software. A developer who uses an AI assistant should review the output, understand its licence implications and check whether it contains code or dependencies that create obligations under third-party licences. The fact that code was generated by an AI tool does not guarantee that it is free from copyright or licensing restrictions.

For small businesses and startups, uncertainty can affect the commercial value of AI-generated marketing materials, product images, websites and training resources. A company may invest in a campaign without checking whether it has the necessary rights to use the material across advertisements, packaging, social media and international markets. Clear contracts and appropriate review processes can reduce these risks.

Practical Steps Creators Can Take to Protect Their Work

Until the law develops greater clarity, creators should maintain evidence of their involvement in AI-assisted projects. This may include early drafts, handwritten sketches, source files, prompt histories, reference materials, editing records and different versions of the final work. Such records can help demonstrate how the project developed and identify the parts created or substantially modified by a human.

Creators should also read the terms of service of the AI platforms they use. These terms may explain whether users receive permissions to use generated outputs commercially, whether the platform retains rights over inputs or outputs, and whether certain uses are restricted. Platform permission and copyright ownership are not necessarily the same thing. A tool may permit commercial use of an output without guaranteeing that the output is independently protected by copyright.

Contracts should be especially clear when AI is used in commissioned work. Freelancers and agencies should specify whether AI tools may be used, who is responsible for reviewing outputs, what rights are being transferred and whether any third-party assets are included. Businesses should also consider whether a client requires disclosure of AI use or prohibits particular tools.

Creators should avoid uploading confidential client information, unpublished manuscripts, private photographs or proprietary business material to AI platforms without checking the platform’s data practices and obtaining the necessary permission. Copyright is not the only legal issue involved in generative AI. Privacy, confidentiality, data protection and contractual duties may also apply.

Finally, creators should not assume that a copyright notice alone establishes ownership. A notice may communicate a claim, but the underlying question remains whether the work qualifies for protection and who is legally entitled to claim it. Where a project has substantial commercial value or a dispute is developing, advice from a qualified intellectual property professional may be necessary.

What Changes May Be Needed in India’s Copyright Framework?

India’s copyright framework faces the challenge of adapting principles developed for human and computer-assisted creativity to systems capable of producing complex content with limited direct intervention. The objective should not simply be to declare all AI-generated material protected or unprotected. A more useful approach would distinguish between fully automated outputs, AI-assisted human creations and works involving substantial human selection, arrangement or modification.

Clearer guidance could explain how Section 2(d)(vi) applies to modern generative AI, what level of human involvement is relevant and how originality should be assessed. It could also clarify how copyright ownership interacts with platform contracts, commissioned works and collaborative projects. Such guidance would help creators, publishers and businesses make informed decisions without treating every AI-assisted work as legally identical.

The government must also consider the interests of creators whose works are used in AI training. Possible policy approaches include licensing arrangements, transparency obligations, collective remuneration mechanisms and clearly defined exceptions. Any framework would need to balance the protection of creators with innovation, research, competition and access to technology.

India’s policy debate has included proposals concerning the use of copyrighted material for AI training and possible remuneration for rights holders. These discussions should not be confused with settled law. A proposal or working paper does not itself create a new copyright entitlement. The status of any recommendation must be checked against the legislation and rules actually in force.

International developments will also influence India’s approach. Countries are taking different positions on human authorship, AI training and the treatment of computer-generated works. Indian creators working with overseas clients or distributing content internationally may need to consider the copyright rules of more than one jurisdiction. A work’s legal status in India does not necessarily determine how it will be treated elsewhere.

The Future of Creativity and Copyright in India

Artificial intelligence is unlikely to replace human creativity in a simple or uniform way. Instead, it is changing how people develop ideas, experiment with formats, produce drafts and refine finished work. The strongest creative projects may increasingly combine human judgment with machine assistance. This makes it important for copyright law to recognise meaningful human contribution without granting automatic ownership over every result produced by a prompt.

For creators, the future will require more than learning how to operate AI tools. They will need to understand licensing, originality, attribution, contracts and the limits of automated generation. Publishers and businesses will need policies that identify when AI is used, preserve evidence of human contribution and check the legal status of material before commercial distribution.

The 2026 DABUS decision has reinforced that an AI system cannot simply be treated as a legal author under India’s existing framework. At the same time, the broader question of who qualifies as the person who causes a modern AI-generated work to be created remains open to interpretation. The answer may depend on the creative process, the human contribution, the nature of the output and the applicable contractual arrangements.

Conclusion

The question of who owns AI-generated content does not have one universal answer under Indian copyright law. Section 2(d)(vi) of the Copyright Act, 1957 recognises the person who causes a computer-generated work to be created as its author, but applying this provision to modern generative AI raises unresolved questions about creative control, originality and human involvement. The Copyright Office’s 2026 DABUS decision makes clear that an AI system cannot itself be recognised as the legal author, but it does not settle every ownership dispute involving AI-assisted content.

For Indian writers, designers, musicians, developers and businesses, the most practical approach is to use AI as a creative tool while maintaining meaningful human involvement, documenting the production process and reviewing the terms governing each platform. Copyright protection should not be assumed merely because a person entered a prompt or paid for an AI service. As generative technology becomes more deeply embedded in creative industries, India’s legal framework will need to provide greater clarity for both innovation and the people whose work gives creativity its value.

Online Internship with Certificate