AI-GENERATED CREATIVITY AND THE COPYRIGHT ACT, 1957 - RETHINKING THE CONCEPT OF ‘AUTHOR’ AND ‘ORIGINALITY’

-       Dr. Soumi Chatterje^[1]

ABSTRACT

This article dives into the big questions around generative AI and copyright law, especially when it comes to ‘authorship’ and ‘originality’. The Indian Copyright Act, 1957, was drafted long before AI came into existence and it presumes that every creative work is produced by a human. Today, advanced language models, image generators, and other AI systems are capable of producing text, code, and artistic works that closely resemble human-created outputs. This development raises significant legal questions that require careful examination: Are these AI-made pieces really ‘original’? Do our laws actually cover things built without a human touch? A related question is whether individuals who guide or prompt AI systems may be regarded as the authors of such outputs. This article explores these issues, comparing Indian law to what’s happening in places like the US and UK. It looks at major cases, Eastern Book Company v. D.B. Modak, Thaler v. Perlmutter, the UK’s Copyright, Designs and Patents Act 1988, and the CJEU’s Infopaq decision. Across the board, courts seem pretty wary about giving copyright protection to works made entirely by machines. Indian law might be able to handle AI-assisted creativity, but fully autonomous AI outputs just don’t fit into the current rules. Accordingly, this article proposes several potential approaches to address these gaps.

Keywords: Artificial Intelligence, Copyright Law, Authorship, Originality, Intellectual Property Policy.

INTRODUCTION

Generative AI systems, including language models, diffusion-based image generators, and autonomous composition algorithms, are transforming the creation and dissemination of creative works. This is reshaping the innovation and distribution of creative works. These technologies can produce text, images, code, and multimedia content at a large scale and velocity which surpass any human capacity. The legal discussions that surround these systems are now centered on a key question: Can AI be considered an author, and if not, who owns these creations?^[2] Indian copyright law, as stated in the Copyright Act, 1957, remains rooted in a human-centric model of authorship and does not explicitly account for creations produced predominantly or exclusively by machines.^[3] Existing statutory definitions accept human creativity as the basis for ownership and protection, creating ambiguity and doctrinal gaps when confronted with the outputs of AI Generated autonomous systems.^[4] This gap is quite severe when compared with international developments. For example, in the UK and the EU, courts and policymakers have begun to hustle with definitions of ‘computer-generated works,’ though ownership remains anchored to human inputs.^[5] In the United States, works lacking sufficient human authorship are generally excluded from copyright protection.^[6] In India, scholarship and practice have yet to converge on a coherent approach, leaving creators, innovators, and legal practitioners in uncertain territory.^[7]

LITERATURE REVIEW

Recent scholarship increasingly examines questions of ownership in relation to AI-generated creative works. As machine learning spits out more and more art, stories, and music, copyright experts are zeroing in on what counts as “original”. The law says a work needs at least a little creativity, and it’s always assumed a human came up with it. Now, with AI in the mix, that old assumption does not feel so simple anymore.^[8] Amankwah-Amoah and others say the old legal system is really struggling to keep up with generative models that churn out content without much human involvement.^[9] Al-Busaidi and his colleagues empirically demonstrate, with real data, that generative AI doesn’t just copy existing work, it can twist and reshape it too. This substantially complicates the determination of copyright ownership.^[10]

Regulators and courts in the US, UK, and India aren’t quick to grant copyright to works that lack enough human input.^[11] Gervais points out that without a human mind behind AI creations, the basic logic of copyright law just doesn’t fit.^[12] But Abbott takes a different route, arguing for a new, custom-made framework that would handle AI-generated works and encourage more of them.^[13] In India, Rasheed and Paul dig into the tricky mix of law and society when it comes to figuring out who counts as an author, and they push for changes in both policy and legal doctrine.^[14]

Kirk et al., dug deeper into how people react to AI taking credit as an author. Studies indicate that attributing authorship to AI systems often generates discomfort and ethical concerns among individuals, people often feel uneasy about it and start questioning what is fair or right. Such attribution also disrupts conventional understandings of credit and responsibility.^[15] On top of that, Lu argues we really need clear rules to sort out who did what the AI or the humans behind it. He pushes for lawmakers to update the rules so they actually fit today’s tech.^[16]

DOCTRINAL ANALYSIS UNDER INDIAN COPYRIGHT LAW

a)    Statutory Definitions and Their Limitations

The Copyright Act of 1957 does not use a one-size-fits-all definition for ‘author’. Usually, it assumes there’s a human behind the creation. However, the application of these definitions to computer-generated works presents significant challenges.^[17] The law says the author is the “person who causes the work to be created.” That made sense back when computers just followed strict instructions and spit out predictable results.^[18] The increasing autonomy of AI systems further complicates this framework. These systems can produce unpredictable stuff that is not directly shaped by a programmer or user, so it is tough to pin down who actually caused the work. On top of that, the law does not even spell out what “original” means. Indian courts have bounced between different tests sometimes they focus on skill, effort, and judgment, other times they look for creativity. The determination of originality varies depending on the specific legal context and judicial interpretation

Categories of AI-Generated Works and Their Legal Status in India

Category	Description	Protectable?	Rationale
A)   Fully Autonomous AI Works	AI generates output without human input post-prompting	No	Lacks human authorship
B)   Human-AI Collaborative Works	AI generates work, human edits, curates, modifies	Possibly	Human contributes creativity
C)   AI-Assisted Works	AI enhances or supports human-authored content	Yes	Human creative control present
D)    Human Work Trained on AI Insights	Human creates original work inspired by AI	Yes	Author is human

                                                                                   Table no. 1

b)    Judicial Treatment of Originality

In Eastern Book Company v. D.B. Modak, the Supreme Court went with a “modicum of creativity” standard, a lot like what the U.S. uses in the Feist case. The Court emphasised that originality requires a minimal degree of human creativity and must originate from a human author.^[19] When it comes to AI-generated stuff, which relies on algorithms and probabilities instead of real human decisions, it probably doesn’t meet that bar.

Similarly, in Academy of General Education v. B. Malini^[20], the Karnataka High Court reaffirmed that The Karnataka High Court made it clear: originality comes from human skill and judgment. To date, Indian courts have not recognised non-human entities as authors for the purposes of copyright protection.

c)     Can AI Be an Author Under Indian Law?

Section 2(d) focuses on humans, and originality still means the creations need a human touch. Further, Indian law doesn’t recognize AI as a legal person. So, AI just cannot be an author until the laws are amended. But here’s where things get disputable, who actually owns work created by AI? The law says ownership goes to the “person who causes the work to be created,” which could mean the programmer, the user, or maybe even the platform. There remains no definitive legal position, as Indian courts have not yet adjudicated upon this question.

COMPARATIVE JURISPRUDENCE

a)    United States

The U.S. Copyright Office doesn’t grant copyright to works made by anything other than people. In Thaler v. Perlmutter, the court backed up the Office’s decision to deny copyright to a piece created by AI, pointing out again that only humans can be authors under U.S. law. ^[21] The Copyright Office’s 2023 Guidance keeps this rule clear that any work of art or creations need real human creativity, and authors must say if they have used AI during their application. Thus, U.S. copyright law firmly restricts authorship to human creators, without exception.

b)    United Kingdom

The UK Copyright, Designs and Patents Act (CDPA) stands out because it actually talks about computer-generated works. In Section 9(3), it calls the author “the person by whom the arrangements necessary for the creation of the work are undertaken.” That’s very similar to India’s Section 2(d)(vi). But UK courts have not really dug into what this means when it comes to autonomous AI. Many scholars contend that the statutory language was drafted with traditional rule-based systems in mind rather than modern generative AI models.

c)     European Union

The CJEU says copyright only covers something that is “the author’s own intellectual creation”, cases like Infopaq, Painer, and BSA make that clear.^[22] So, if a work comes entirely from AI with no human in the loop, it doesn’t qualify for copyright. Right now, the EU’s AI Act proposals talk a lot about transparency and how to handle datasets, but they do not touch on who gets to be called an author. Back in 2020, the European Parliament clearly said no to giving AI legal personhood.^[23]

d)    Comparative Observations

Across jurisdictions:

• AI is not recognised as an author.

• Copyright requires human creativity.

• Users or developers may claim ownership only where they make creative choices.

The trend is clear: human creativity is indispensable for copyright protection.

POLICY ANALYSIS

a)    Problems with Applying Existing Law to AI-Generated Works

1. Ambiguity of Authorship: Determining the appropriate author in the context of AI-generated works is particularly complex. Programmers write the code, dataset curators pick the data, prompt engineers shape the instructions, and end-users actually use the system. Because multiple actors contribute at different stages, establishing authorship becomes difficult

2. Over-monopolisation by technology companies: Attributing authorship to developers may risk concentrating excessive control in the hands of large technology companies. They already have most of the resources, and these just hands them even more control.

3. Dilution of the originality standard: If we start giving copyright to works made entirely by machines, the creative space could get flooded with stuff that doesn’t involve any human creativity. This risks lowering the standard for what constitutes originality

4. Moral rights incompatibility: AI cannot claim personal rights like integrity or attribution. Those rights are meant for real people, not machines.

b)    Arguments for Reform

1. Economic Incentives

   Developers want clear rights before they’ll put serious money into something.

   
   

2. Sui- Generis Protection

   Giving a short-term right, kind of like what we do for databases, can encourage investment without locking up access forever.

   
   

3. Contractual Allocation

   Platforms often lean on licenses instead of full-blown copyright laws. But honestly, this just leads to a messy patchwork that’s tough to navigate.

   
   

4. Mandatory Disclosure

   If a work uses AI, maybe we should just say so and spell out how much of it came from a human.

c)     Aligning Indian Law with Innovation Policy

India’s National IPR Policy (2016) talks a lot about innovation and new tech, but it doesn’t really say what to do about AI authorship. India requires a targeted strategy that works in order to advance. Fundamentally, the law should safeguard human creativity, specify who owns works supported by AI, prevent large corporations from obtaining all the rights, and ensure that science and creativity are accessible to everybody.

d)    Comparative Authorship Standards Under Major Copyright Regimes

Jurisdiction	Statutory Basis	Definition of Author	Position on AI as Author	Leading Case / Authority
India	Copyright Act 1957, Sec. 2(d)	Human person who creates the work	Not allowed -- human authorship required	EBC v. D.B. Modak (2008)
US	Copyright Act 1976	Work must be created by a human being	Not allowed -- AI cannot hold copyright	Thaler v. Perlmutter (2023)
UK	CDPA 1988, Sec. 9(3)	Person making necessary arrangements	Allows authorship of “computer-generated works”	--
EU	InfoSoc Directive; CJEU jurisprudence	Requires author’s “own intellectual creation”	Not allowed -- strictly human	Infopaq (2009)

                                                                       Table No.2

e)     A Comparative Analysis of Tests of Originality

Jurisdiction	Standard of Originality	Human Involvement Required	Applicability to AI Outputs
India	Modicum of creativity (Modak)	Yes	Limited to AI-assisted works
US	Minimal creativity (Feist)	Yes	AI-autonomous work excluded
UK	Skill, labour and judgment	Yes (but broader)	Partial recognition via CDPA §9(3)
EU	Intellectual creation of author	Strictly yes	Not applicable to autonomous AI

                                                                              Table No. 3

CONCLUSION

AI-generated creativity exposes significant gaps within existing copyright law, particularly relating to authorship and originality. The Copyright Act of 1957 in India dates back to a period when no one could have imagined machines creating art or narratives. It’s all built around the idea that copyright protects what people create. If we look at the US, UK, and EU, they’re on the same page as humans are at the center of authorship. Writers and legal experts mostly agree that AI itself can’t be an author. But when we ask if or how we should protect works made by AI, that’s when confusion or ambiguity arise. People argue about ownership, worry about turning everything into just another product, and point out that current rules don’t fit. Further, AI-generated work does not really match up with how we think about moral rights. All of this makes it very clear that copyright law needs to be updated and amended as per contemporary societal and creative needs. The key issue, therefore, is determining the most appropriate course of action for India in light of these developments. Determine who gets credit for AI-assisted works, reject copyright for entirely AI-generated work, establish a new type of right that only applies to AI-generated work, require disclosure of AI use, define fair use guidelines, and safeguard the public domain. India can keep up with the rest of the world, safeguard human ingenuity, and ensure that the general public continues to gain if it can make such corrections. The real future of copyright is not about striking out human authors, but about figuring out how people and machines can create together and what that implies for everyone.

^[1] Dr. Soumi Chatterjee is the Founder/CEO of SRTD Group & Vidhi Vaarta.

^[2] Khurana & Khurana, Ownership of AI Generated Content: A Deep Dive into Copyright Law in India, (Oct. 24, 2025), Available at https://www.khuranaandkhurana.com/ownership-of-ai-generated-content-a-deep-dive-into-copyright-law-in-india (last visited Nov. 15, 2025).

^[3] Avishek Chakraborty, Authorship of AI Generated Works under the Copyright Act, 1957:An Analytical Study, Nirma University Law Journal, Vol 8, Issue 2 (2019), Available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3443006 (last visited Nov. 15, 2025).

^[4] Abdullah S. Al-Busaidi et al., Investigating the Impact of Generative Artificial Intelligence on the Replication and Transformation of Copyrighted Works, 42 J. Info. Sci. 1181 (2024), Available at https://www.sciencedirect.com/science/article/pii/S2444569X24001690 (last visited Nov. 15, 2025).

^[5] Emil Dai, A Comparative Study of Copyright Protection for AI-Generated Works in the US, UK and EU: Exploring the Legal Status of AI as Author, (Nov. 4, 2024), Available at https://emildai.eu/publication/a-comparative-study-of-copyright-protection-for-ai-generated-works-in-the-us-uk-and-eu-exploring- (last visited Nov. 16, 2025).

^[6]Prashant Shukla & Dr. Mudra Singh, Generative AI And Copyright Law: A Comparative Analysis of Indian and International Frameworks, IJLLR (Apr. 28, 2025), Available at https://www.ijllr.com/post/generative-ai-and-copyright-law-a-comparative-analysis-of-indian-and-international-frameworks  (last visited Nov. 17, 2025).

^[7] Jasneet Kaur, Artificial Intelligence and Protection of AI Generated works under Indian Copyright Law, 3 ShodhKosh J. Arts, Human. & Soc. Sci. 1 (2024), Available at https://www.granthaalayahpublication.org/Arts-Journal/ShodhKosh/article/view/5469  (last visited Nov. 17, 2025).

^[8] Ploman Edward & Lenn Hamilton, Copyright, Originality and Authorship: Reconciling Legal Standards in the Age of AI, 9 L.J. DIGITAL 33, 35 (2024).

^[9] John Amankwah-Amoah et al., The Impending Disruption of Creative Industries by Generative Artificial Intelligence: Implications and Opportunities, 75 TECHNOL. FORECAST. SOC. CHANGE 103112 (2024).

^[10] Abdullah S. Al-Busaidi et al., Investigating the impact of generative artificial intelligence on the replication and transformation of copyrighted works, 42 J. INFO. SCI. 1181 (2024).

^[11] Daniel Gervais, The Machine as Author, 105 IOWA L. REV. 2053, 2055 (2020).

^[12] Id.

^[13] Ryan Abbott, I Think, Therefore I Invent: Creative Computers and the Future of Patent Law, 57 B.C.L. REV. 1079, 1081 (2016).

^[14] M. Rasheed & J. Paul, Artificial Intelligence and Copyright in India: A Socio-Legal Analysis, 12 IND. J. L. & TECH. 60, 64 (2024).

^[15] Colby P. Kirk et al., The AI-authorship effect: Understanding authenticity, moral disgust, and loyalty to creative works, 143 J. BUS. RES. 481, 485 (2025).

^[16] Yang Lu, Reforming Copyright Law for AI-Generated Content, 44 INT’L J. LAW & INFO. TECH. 50, 53 (2025).

^[17] Copyright Act, 1957, Sec. 2(d).

^[18] Id. Sec. 2(d)(vi).

^[19] Eastern Book Co. v. D.B. Modak, (2008) 1 SCC 1.

^[20] AIR 2009 SC 1982.

^[21] Thaler v. Perlmutter, 2023 WL 5333236 (D.D.C. 2023).

^[22] Infopaq International A/S v. Danske Dagblades Forening, Case C-5/08 (CJEU 2009); Painer, Case C-145/10; BSA, Case C-393/09.

^[23] See European Parliament Res., 2020/2014(INL), Resolution on a Civil Liability Regime for Artificial Intelligence (July 16, 2020), Available at https://www.europarl.europa.eu/doceo/document/TA-9-2020-0214_EN.html (last visited Nov. 17, 2025); see also European Parliament Res., 2020/2015(INI), Resolution on Intellectual Property Rights for Artificial Intelligence Development (July 16, 2020), Available at https://www.europarl.europa.eu/doceo/document/TA-9-2020-0215_EN.html (last visited Nov. 17, 2025); see also generally Refusing to Award Legal Personality to AI, European Law Blog (Nov. 24, 2020),Available at https://www.europeanlawblog.eu/pub/refusing-to-award-legal-personality-to-ai-why-the-european-parliament-got-it-wrong (last visited Nov. 17, 2025).