AI-Assisted vs. Fully AI-Generated Works – Where Indian Copyright Law Draws the Line

The emergence of generative artificial intelligence as a practical tool for creative production has forced copyright law to confront a question it was never designed to answer: at what point does a work cease to be the product of human creative expression and become the output of a machine? This question matters enormously, for the artists, writers, musicians and software developers who use AI tools in their creative workflows, for the businesses that commission and deploy AI-generated content, for the platforms that distribute it and for the courts and legislators who must determine whether and how the law’s protection extends to it.

The question is not binary. The creative use of artificial intelligence exists on a spectrum that ranges from the barely perceptible – a novelist who uses a grammar-checking tool to refine their prose to the almost entirely automated, a user who types a five-word prompt and accepts the AI’s unmodified output as their finished work.Between these poles lies an enormous range of human-AI collaborative arrangements, each presenting a different factual picture of who did what in the creative process and whether the human’s contribution crosses the threshold that copyright law requires.

Indian copyright law does not yet have a clear judicial or legislative answer to most of the specific questions raised by this spectrum. What it does have is a statutory framework the Copyright Act, 1957, as amended, whose provisions can be analyzed and applied to these questions with varying degrees of confidence. Section 2(d)(vi), the provision expressly addressing computer-generated works, provides a starting point. The originality requirement established in Eastern Book Company v. D.B. Modak supplies the doctrinal test. The idea-expression dichotomy defines the outer boundaries of what copyright can protect in any work, however created. And the comparative experience of other jurisdictions, particularly the developing jurisprudence of the United States Copyright Office and the emerging case law of American federal courts provides contextual reference points that Indian decision-makers are likely to consider.

Setting the Stage – The AI Creative Spectrum

To understand where the law draws the line, it is necessary first to map the terrain across which that line must be drawn. AI-assisted creation and fully AI-generated creation are not two sharply distinct categories but points on a continuum defined by the degree of human creative contribution relative to the AI system’s autonomous creative role.

At one end of the spectrum is the human creator who uses AI as a passive instrument – a tool that executes the creator’s specific, detailed creative instructions without exercising any independent creative judgment. The author who uses a word processor, the composer who uses notation software, the graphic designer who uses image editing software, all use computational tools that facilitate the expression of human creative vision without themselves making creative choices. These uses are so clearly within the domain of traditional human authorship that they raise no copyright question. The tool executes; the human creates.

Moving along the spectrum, we encounter the creator who uses AI as an intelligent assistant – a tool that exercises limited autonomous creative judgment within a framework strongly defined by the human’s creative direction. The author who uses an AI writing tool to generate candidate sentences from which they select and edit, the designer who uses an AI layout tool to generate composition options from which they choose, the musician who uses an AI harmonisation tool to suggest chord progressions that they accept, modify or reject – all of these creators are directing an AI system that exercises some creative judgment, but the human’s contribution remains dominant, specific and clearly expressive.

Further along the spectrum is the creator who uses AI as an active creative collaborator – providing general creative direction and parameters, making selections among AI-generated alternatives and curating and refining the AI’s output into a finished work. The filmmaker who uses AI-generated visual effects, the writer who uses an AI system to generate plot options from which they select and develop, the graphic artist who uses a diffusion model to generate image components that they assemble and compose – these creators exercise meaningful creative judgment, but the AI system’s autonomous creative contribution to the final work is substantial.

At the far end of the spectrum is the person who uses AI as an autonomous creative agent providing a minimal prompt and accepting the AI’s output without significant modification. The user who types “write a short story about a dog” and publishes the AI’s unmodified response, the user who types “paint a landscape in the style of impressionism” and submits the diffusion model’s first output as their artwork. these users have exercised minimal creative judgment about the work’s specific expressive content. The conceptual direction is human; the creative execution is entirely the machine’s.

Indian copyright law must determine, across this spectrum, at what point human creative contribution is sufficient to attract copyright protection and in whom that protection vests.

The Statutory Framework – Section 2(d)(vi) and Its Limits

The starting point for any analysis of AI-generated works under Indian copyright law is Section 2(d)(vi) of the Copyright Act, 1957, which provides that in the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be the person who causes the work to be created.

This provision is, in the international context, unusual in its directness. Most copyright jurisdictions have no express statutory provision addressing the authorship of computer-generated works. The United Kingdom has Section 9(3) of the Copyright, Designs and Patents Act, 1988, which takes the same approach as India’s Section 2(d)(vi). Most other jurisdictions including the United States, the member states of the European Union and most Asian copyright systems have no equivalent provision and address the question through the general originality and authorship framework.

India’s Section 2(d)(vi) performs two functions. First, it recognises that copyright can subsist in computer-generated works – works produced by a computational process without a conventional human author making the specific expressive choices. This recognition is significant because without it, the application of the general originality requirement to computer-generated works would likely produce the conclusion that such works attract no copyright at all, since there is no human author whose intellectual creation the work embodies. Second, it identifies the author of such a work as the person who causes it to be created providing a rule for the allocation of authorship and therefore of first ownership in works that lack a conventional human creator.

The critical limitation of Section 2(d)(vi) is that it was enacted in 1994 a quarter century before the current generation of large language models, diffusion models and other generative AI systems came into existence. Its legislative context was the modest forms of computer-assisted creation prevalent in the early 1990s: algorithmic composition tools, report-generating programs, database output systems. The provision was designed to ensure that works generated by deterministic computational processes – programs that transform inputs into outputs according to fixed rules could attract copyright and vest that copyright in the person who set the process in motion.

Modern generative AI systems are fundamentally different from the “computer-generated” works that Section 2(d)(vi)’s drafters had in mind. Large language models like GPT-4 and Claude generate outputs by sampling from probability distributions over possible continuations of a given input, producing responses that are statistically plausible rather than algorithmically determined. Diffusion models like Stable Diffusion and Midjourney generate images by a process of iterative denoising that produces outputs sensitive to the latent statistical patterns learned from billions of training examples. These systems exercise something that, while not human creative judgment, is meaningfully different from the execution of fixed rules – they make choices, in a statistical sense, that determine the specific character of their outputs in ways that are not fully determined by the input prompt.

The question of whether Section 2(d)(vi)’s phrase “person who causes the work to be created” adequately captures the human-AI relationship in modern generative AI usage is therefore genuinely uncertain. The provision can be applied to modern AI contexts, but its application requires interpretation that the drafters did not undertake and that courts have not yet provided

The Originality Requirement – The Doctrinal Core

The second pillar of the legal analysis is the originality requirement. Section 13(1)(a) of the Copyright Act provides that copyright shall subsist in original literary, dramatic, musical and artistic works. The requirement of originality is not defined in the Act but has been extensively analysed by Indian courts, with the Supreme Court’s decision in Eastern Book Company v. D.B. Modak (2008) establishing the applicable standard.

Eastern Book Company adopted what the Supreme Court described as a “minimal degree of creativity” standard – not mere labour, not mere effort, but the exercise of some creative judgment by the author that reflects their own intellectual creation. The work must originate from the author, it must be the product of the author’s creative choices and it must involve some element of creativity, however modest, beyond purely mechanical or algorithmic production.

Applied to the AI creative spectrum, the originality requirement produces the following analysis.

For works created with AI as a passive instrument where the AI executes the human’s specific creative instructions without exercising independent creative judgment, the originality analysis is straightforward. The human author’s creative choices determine the expressive character of the work. The AI tool’s role is purely instrumental, equivalent to the role of a paintbrush, a word processor or a camera. The work satisfies the originality requirement in the human author, who is the author for all copyright purposes.

For works created with AI as an intelligent assistant, where the AI exercises limited autonomous creative judgment within a strongly human-defined framework, the originality analysis requires a more careful assessment of the human’s creative contribution. Where the human’s creative direction is specific and detailed providing not merely a general topic but specific expressive choices about structure, style, content and form, the human’s contribution likely satisfies the originality requirement. Where the human’s direction is general and the AI’s specific expressive choices are substantial, the analysis becomes less certain.

For works created with AI as an active creative collaborator where the human provides general direction but the AI makes substantial specific expressive choices — the originality analysis is genuinely uncertain. The human has made creative choices about the parameters of the creative brief, about which AI-generated alternatives to select, about how to assemble and refine the AI’s outputs but the most specific expressive choices determining the work’s character may have been made by the AI system. Whether the human’s contribution constitutes a sufficient “minimal degree of creativity” to attract copyright in the finished work is a question that Indian courts have not yet answered.

For works created with AI as an autonomous creative agent, where the human provides a minimal prompt and accepts the AI’s output without significant modification, the originality analysis is closest to the position of the United States Copyright Office, which has consistently held that works generated by AI without sufficient human creative contribution do not attract copyright. If the human has made no specific expressive choices about the work’s content – if they have specified only the most general category of output desired and the AI has made all the specific choices that give the work its expressive character, it is difficult to see how the originality requirement is satisfied by the human’s contribution.

The “Causes to Be Created” Test – What Does It Require?

The operative language of Section 2(d)(vi) that the author of a computer-generated work is “the person who causes the work to be created” is the primary legal test for authorship of AI-generated works under Indian law. Understanding what this test requires and how it applies across the AI creative spectrum, is essential to determining where Indian copyright law draws the line.

The phrase “causes the work to be created” describes a causal relationship between a person and a work. At its most expansive, virtually any person who interacts with an AI system to produce a work “causes” that work without their input, the work would not exist. On this maximally expansive reading, the user who types a single-word prompt and accepts the AI’s output is the “person who causes” the resulting work to be created.

At its most restrictive, the causal relationship required by Section 2(d)(vi) might be interpreted as requiring that the person’s specific creative choices be the operative cause of the work’s specific expressive character not merely the proximate cause of any output being generated at all. On this restrictive reading, a person who provides an extremely general direction to an AI system and accepts whatever it produces without modification has caused something to be created, but has not caused the specific work that resulted to be created in any meaningful sense, the AI’s autonomous choices determined the work’s specific character and the human’s contribution was the general occasion rather than the specific cause.

The most legally defensible interpretation of Section 2(d)(vi) in the context of modern generative AI is an intermediate one that requires the human’s interaction with the AI system to involve some meaningful degree of creative direction. some exercise of the person’s own creative judgment in specifying, selecting or refining the AI’s output before they can be said to be the “person who causes the work to be created” within the meaning of the provision. On this interpretation, the provision is satisfied where the human’s creative direction is sufficient to make their contribution the expressive cause of the work’s distinctive character, but is not satisfied where the human’s direction is so general that the AI’s autonomous choices entirely determine that character.

This interpretation aligns the “causes to be created” test with the originality requirement both tests point toward the same conclusion that human creative contribution of some meaningful degree is necessary for copyright to subsist in the human prompter. Where the human contribution satisfies the “causes to be created” test, it likely also satisfies the originality requirement; where it does not satisfy the originality requirement, it likely also falls short of “causing” the work to be created in the relevant sense.

How Indian Courts Are Likely to Approach the Spectrum

While no Indian court has directly ruled on the copyright status of AI-generated works, the existing jurisprudence provides a basis for predicting how courts are likely to approach the spectrum of human-AI creative collaboration.

For works at the human-dominant end of the spectrum where AI is used as a tool for the execution of the human’s creative vision Indian courts are likely to apply conventional copyright analysis and find that copyright subsists in the human author whose creative choices determine the work’s expressive character. The fact that a computational tool was used in the creation process does not, on its own, displace human authorship. Indian courts have consistently applied a functional analysis of creative contribution rather than a formalistic analysis of the means by which creation occurred.

For works in the middle of the spectrum, where human and AI contributions are meaningfully intertwined, Indian courts are likely to conduct a fact-specific analysis of the human’s creative contribution, asking whether that contribution satisfies the “minimal degree of creativity” standard of Eastern Book Company. Courts are likely to consider the specificity of the human’s creative direction, the nature and extent of the human’s selection and curation of AI outputs, the degree of human editing and modification applied to AI-generated material and the extent to which the finished work reflects the human’s own expressive choices rather than the AI’s statistical choices.

For works at the AI-dominant end of the spectrum, where the human provides minimal creative direction and accepts AI output without significant modification – Indian courts are likely to find that the originality and “causes to be created” requirements are not satisfied and that no copyright subsists in the work as a result of the human’s minimal contribution. This outcome aligns with the position taken by the United States Copyright Office, the English Court’s approach in the Thaler litigation and the general direction of emerging international jurisprudence on AI-generated works.

The Prompt as Creative Expression – A Critical Assessment

A significant argument advanced in support of copyright protection for AI-generated outputs is that the human’s prompt, the instruction or description provided to the AI system is itself a form of creative expression that translates into copyright in the resulting work. On this argument, a carefully crafted prompt that specifies detailed creative parameters represents the human’s artistic vision and the AI’s output is the expression of that vision, analogous to the execution of a commissioned artwork based on the patron’s creative brief.

The prompt-as-creative-expression argument has intuitive appeal in cases where prompts are themselves detailed, specific and creatively sophisticated, where the prompt specifies not just the general subject but the specific aesthetic approach, compositional elements, stylistic references, emotional register and expressive qualities desired in the output. A prompt of this character represents a genuine creative contribution by its author and there is a plausible argument that the resulting AI output translates that creative contribution into expressive form.

However, the argument has significant limitations. The connection between the human’s prompt and the AI’s specific expressive choices is not the same as the connection between a human author’s intentions and their own creative execution. When a human author writes a sentence, the specific words, syntactic choices and expressive details are all products of the author’s own creative judgment. When an AI system generates a sentence in response to a prompt, the specific words, syntactic choices and expressive details are products of the AI’s statistical sampling process they reflect the prompt’s general direction but are determined by the AI’s learned patterns rather than the human’s specific creative choices.

The United States Copyright Office has specifically rejected the argument that a detailed prompt gives rise to copyright in the AI’s output, noting that the critical question is not whether the human specified what they wanted but whether the human’s own creative expression appears in the work. A prompt specifies desired outcomes; it does not itself determine the specific expressive choices that constitute the work. The AI’s specific choices the particular words of a poem, the specific compositional arrangement of an image, the distinctive harmonic progression of a musical piece — are the AI’s choices, not the prompter’s, regardless of how detailed the prompt was.

Indian courts applying the Eastern Book Company originality standard are likely to reach a similar conclusion. The minimal creativity required by that standard must be creativity expressed in the work itself it must be the human’s own intellectual creation that appears in the copyrightable expression. A prompt that specifies desired outcomes without determining the specific expressive choices that constitute the work does not satisfy this requirement, regardless of the prompt’s sophistication or the care with which it was crafted.

The Selection, Arrangement and Curation Argument

A more promising basis for copyright protection in AI-assisted works is the selection, arrangement and curation argument the argument that where a human creator exercises meaningful creative judgment in selecting from among AI-generated alternatives, arranging AI-generated components and curating AI outputs into a finished work, those selection and arrangement choices constitute original creative expression protectable by copyright.

Copyright in compilations and arrangements has long been recognised as a distinct form of copyright protection the creative selection and arrangement of pre-existing materials can attract copyright independently of any copyright in the materials selected and arranged. If a human creator uses an AI system to generate a series of candidate images, selects a subset of those images based on their own aesthetic judgment, arranges them in a sequence or composition reflecting their creative vision and adds their own expressive elements to the arrangement, the resulting work may attract copyright in the selection and arrangement regardless of the copyright status of the individual AI-generated components.

This argument is most compelling where the creative selection and arrangement choices are themselves numerous and specific where the human has made many individual creative decisions about which elements to include, how to arrange them and what relationships to create between them. It is less compelling where the selection involves choosing a single output from a small number of AI-generated alternatives or where the arrangement is dictated by obvious or mechanical principles rather than by the human’s own creative judgment.

The Eastern Book Company case itself involved exactly this type of selection and arrangement analysis the Supreme Court examined which editorial additions to judicial decisions reflected creative selection choices by the editor and which were mechanical. The principles established in that analysis provide direct guidance for the assessment of selection and arrangement creativity in the AI context.

The Derivative Work Question – When AI Output Builds on Human Creative Input

A related question is the copyright status of AI outputs that build substantially on human-created works provided as inputs. Where a human creator provides their own original creative work as input to an AI system a photographer who uploads their original photographs to an AI-powered style transfer tool, a writer who provides their original manuscript to an AI editing and expansion tool, a musician who provides their original composition to an AI arrangement tool the AI’s output may be derivative of the human’s original work.

In such cases, the copyright analysis is more favourable to the human creator. The AI-generated output, to the extent that it reproduces or derives from the human’s original creative expression, is a derivative work that builds on the human’s copyright. The adaptation right under Section 14(a)(vi) would be engaged by the AI’s transformation of the human’s work and the human as copyright owner of the underlying work has the exclusive right to authorize such adaptations.

The human’s own expressive contribution the original work provided as input is clearly protectable by copyright regardless of the AI’s involvement. The question is whether the AI’s transformation adds sufficient additional original expression to attract independent copyright in the derivative work. Where the AI’s transformation preserves and builds upon the human’s original creative choices, the resulting derivative work may attract copyright in the combination of the human’s underlying expression and the human’s direction of the transformation, even where the specific added elements were generated by the AI.

This analysis has practical implications for the emerging field of AI-assisted creative development where creators use AI tools to explore variations on their original creative concepts, to generate derivative versions of their original works for different contexts or audiences or to develop their creative ideas beyond what their own technical skills would allow. These uses may be distinguished from purely AI-generated creation on the ground that the human’s original creative expression forms the foundation of the resulting work.

Comparative Framework – What Other Jurisdictions Tell Us

The comparative experience of other jurisdictions in addressing the AI creative spectrum provides important reference points for the development of Indian copyright doctrine, even where those jurisdictions’ specific legal frameworks differ from India’s.

The United States Copyright Office has developed the most detailed framework for assessing copyright in AI-generated works through its registration practice and published guidance. The Office has articulated a standard based on whether the human author has made “sufficient creative control” over the work whether the human’s own expressive choices are discernible in the final product. In Zarya of the Dawn (2023), the Office registered copyright in a graphic novel whose text was human-authored and whose overall selection and arrangement of images was human-curated, while refusing registration for the individual AI-generated images on the grounds that they were not products of human creative expression. This nuanced approach recognizing copyright in the human-curated whole while declining to protect the AI-generated components is a model that Indian courts might find persuasive.

The English law approach, as expressed in Section 9(3) of the CDPA and the emerging case law, has historically been more generous to computer-generated works than the American approach, recognizing copyright in works generated by computers without requiring human authorship in the creative sense. However, the English courts have not yet directly addressed the specific question of modern generative AI works and the adequacy of the existing framework for these purposes is a subject of active academic and policy debate.

The Court of Justice of the European Union has, through decisions on copyright in photographs and other borderline works, established that the European copyright standard requires that the work reflect the author’s “own intellectual creation” that it reflect the author’s free and creative choices. This standard, applied to AI-generated works, requires that the specific expressive choices constituting the work be the human author’s own choices, not the AI’s — a requirement that will not be satisfied by works whose specific expressive character was determined by the AI’s autonomous creative choices.

The Chinese approach has been the most permissive in the jurisdiction so far the Beijing Internet Court in Li Yunkai v. Liu Yuanchun (2023) recognized copyright in an AI-generated image on the basis that the human plaintiff’s iterative prompting and selection reflected sufficient creative effort. This decision remains controversial and its precedential scope is uncertain, but it demonstrates that some jurisdictions are considering a more expansive approach to AI authorship that focuses on the overall creative process rather than the moment of specific expressive choice.

Practical Implications for Creators and Businesses

The current state of Indian copyright law on the AI creative spectrum uncertain in many respects, but analyzable through the available statutory and doctrinal framework has several practical implications for creators and businesses using AI tools in their creative work.

For creators who use AI as a tool within a strongly human-directed creative process, the copyright position is relatively clear. Works in which the human’s creative choices are dominant and specific where the AI’s contribution is the execution of detailed human instructions rather than the making of independent creative choice are likely to attract copyright in the human creator under conventional copyright principles. These creators should document their creative process to establish the human contribution’s specificity and dominance in case of any future dispute.

For creators who use AI as a creative collaborator providing general direction and making selections among AI-generated alternatives the copyright position is less certain but not hopeless. Copyright may subsist in the human’s creative selections, arrangements and editorial choices, even where the AI’s specific expressive choices are substantial. These creators should maintain records of their creative process, preserve evidence of the creative decisions they made and structure their creative workflow to maxima the identifiable human contribution to the finished work.

For businesses that commission or deploy AI-generated content marketing materials, product imagery, music for advertisements, website copy generated by AI writing tools the copyright uncertainty creates practical challenges for ownership and enforcement. Businesses should ensure that their agreements with AI platform providers address copyright ownership in outputs, should document their creative involvement in AI-generated content to support ownership claims and should be aware that content generated with minimal human creative direction may not be protectable by copyright under Indian law.

For platforms and distributors that host AI-generated content, the uncertainty about copyright subsistence creates challenges for content moderation, licensing and the management of competing claims. A platform that hosts AI-generated works faces potential liability for infringement of training data copyrights, uncertainty about whether the platform users hold enforceable rights in their AI-generated content and the possibility that AI-generated content presented as human-authored will be the subject of misrepresentation claims.

The Registration Question – Should AI-Assisted Works Be Registered?

The Copyright Office of India has not yet issued guidance on whether AI-generated or AI-assisted works may be registered or what information must be disclosed about AI involvement in the creation process. This absence of guidance creates uncertainty for rights holders who wish to register works created with AI assistance.

The Copyright Act’s registration system is voluntary and the registration certificate creates only a prima facie presumption of the facts stated in the register. A person who registers a work and falsely states that it was entirely human-authored when it was substantially AI-generated provides inaccurate information to the Copyright Office, potentially undermining the certificate’s evidentiary value and exposing the registrant to the consequences of misrepresentation.

Pending formal guidance from the Copyright Office, rights holders registering AI-assisted works should consider accurately disclosing the AI’s involvement in the creation process and clearly identifying the human creative contributions that form the basis of the copyright claim. This approach minimises the risk of successful challenge to the registration and provides an honest foundation for the copyright claim.

The United States Copyright Office’s developing practice of accepting registrations for AI-assisted works that involve sufficient human creative contribution while requiring disclosure of AI involvement provides a practical model that the Indian Copyright Office might adopt. Clear guidance on disclosure requirements and the standards for registrability of AI-assisted works would significantly reduce the uncertainty that rights holders currently face.

The Line Indian Copyright Law Draws – A Synthesis

Drawing together the statutory framework, the applicable doctrinal tests, the emerging comparative jurisprudence and the practical implications examined above, it is possible to articulate where Indian copyright law is likely to draw the line between protectable AI-assisted works and unprotectable fully AI-generated works.

Copyright is likely to subsist where the human creator’s contribution satisfies three conditions: the contribution involves specific creative choices about the work’s expressive character rather than merely general direction about desired outcomes; those creative choices are the author’s own intellectual creation rather than the product of the AI’s autonomous statistical processes; and the creative choices constitute at least the “minimal degree of creativity” required by the Eastern Book Company standard.

Copyright is likely not to subsist where the human’s contribution consists only of providing a general prompt, accepting the AI’s output without significant modification and perhaps selecting from a small number of AI-generated alternatives. In such cases, the specific expressive choices that constitute the work’s creative character are the AI’s choices, not the human’s and the originality and “causes to be created” requirements are not satisfied.

Between these poles lies a wide zone of uncertainty cases where the human’s creative direction was moderately specific, where the selection from AI alternatives involved genuine aesthetic judgment, where the human made meaningful editorial choices about AI-generated material or where the human combined AI-generated elements with their own original expression. For works in this zone, the copyright determination will be fact-specific, context-dependent and potentially unpredictable until Indian courts develop a more detailed jurisprudence for the assessment of human-AI creative collaboration.

The Legislative Response – What India Needs

The existing statutory framework, applied through judicial interpretation, can address some of the questions raised by AI-assisted and AI-generated works. It cannot address all of them and the pace of AI’s integration into creative production means that the gap between the framework’s capacity and the questions it must answer will widen unless legislative reform addresses it directly.

India needs, at minimum, three legislative interventions in the copyright domain of AI. First, a clear statutory definition of what “computer-generated” means in the context of modern generative AI one that acknowledges the qualitative difference between deterministic algorithmic output and the statistical creative processes of large language models and diffusion models. Second, clear standards for what level of human creative contribution is necessary to attract copyright in AI-assisted works standards that go beyond the current reliance on judicial interpretation of the “causes to be created” language. Third, a text and data mining exception that provides a legal basis for the training of AI systems on copyrighted works, resolving the training data infringement question that creates legal uncertainty for AI developers and rights holders alike.

These reforms would not resolve every question the creativity and ingenuity of AI developers ensures that new questions will continue to emerge faster than any single legislative intervention can answer them. But they would provide a clearer foundation than currently exists for the assessment of copyright in AI-generated works and would enable Indian creators, businesses and courts to navigate the AI creative landscape with greater legal certainty.

Conclusion

The line between AI-assisted and fully AI-generated works is not a bright line in Indian copyright law it is a zone of gradation whose specific boundaries are defined by the interaction of Section 2(d)(vi)’s “causes to be created” test, the Eastern Book Company originality standard and the idea-expression dichotomy. Works in which the human’s specific creative choices are dominant and clearly expressed in the finished work attract copyright in the human creator. Works in which the AI’s autonomous statistical choices entirely determine the specific expressive character of the output do not or at least cannot confidently be said to attract copyright under the current framework.

For the enormous range of creative works that lie between these poles works that combine meaningful human creative direction with substantial AI creative execution the copyright position remains genuinely uncertain under existing Indian law. That uncertainty is not a failure of the law’s basic principles but a reflection of the unprecedented nature of the creative arrangements that generative AI has made possible. The principles are clear enough; their application to novel facts requires the judicial development that has not yet occurred and the legislative guidance that has not yet been provided.

The practical message for creators and businesses in the interim is to maxima the human creative contribution in AI-assisted workflows, to document that contribution carefully, to be transparent about AI involvement in the registration process and to be aware that works generated with minimal human creative direction occupy a zone of copyright vulnerability that may ultimately require legislative resolution. The creative economy will not wait for the law to catch up but understanding where the law currently stands is the essential foundation for navigating it wisely.

References

1. The Copyright Act, 1957, Sections 2(d)(vi), 2(o), 13, 14, 17, 22, 52, 57 – https://copyright.gov.in/Documents/CopyrightRules1958.pdf
2. The Copyright (Amendment) Act, 2012 – https://copyright.gov.in/Documents/Amendment_Act2012.pdf
3. Eastern Book Company v. D.B. Modak, (2008) 1 SCC 1 (Supreme Court of India) – https://indiankanoon.org/doc/1023365/
4. R.G. Anand v. Deluxe Films & Ors., AIR 1978 SC 1613 (Supreme Court of India) – https://indiankanoon.org/doc/595730/
5. Thaler v. Perlmutter, Case No. 22-1564 (D.D.C. 2023) – https://www.courtlistener.com/docket/63356632/thaler-v-perlmutter/
6. Zarya of the Dawn – U.S. Copyright Office Registration Decision (2023) – https://www.copyright.gov/ai/
7. U.S. Copyright Office — Copyright and Artificial Intelligence Policy Statement (2023) – https://www.copyright.gov/ai/
8. Li Yunkai v. Liu Yuanchun (2023), Beijing Internet Court – http://www.bjinternetcourt.gov.cn
9. UK Copyright, Designs and Patents Act, 1988, Section 9(3) – https://www.legislation.gov.uk/ukpga/1988/48/contents
10. EU Directive on Copyright in the Digital Single Market, 2019/790 – https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32019L0790
11. WIPO – Conversation on Intellectual Property and Artificial Intelligence – https://www.wipo.int/about-ip/en/artificial_intelligence/
12. Berne Convention for the Protection of Literary and Artistic Works, Article 2 – https://www.wipo.int/treaties/en/ip/berne/
13. TRIPS Agreement – https://www.wto.org/english/docs_e/legal_e/27-trips.pdf
14. Copyright Office of India – https://copyright.gov.in
15. Amarnath Sehgal v. Union of India, 117 (2005) DLT 717 (Delhi High Court) – https://indiankanoon.org/doc/1402532/

Frequently Asked Questions: