ARTIFICIAL INTELLIGENCE AND COPYRIGHT: NAVIGATING OWNERSHIP OVER AI-GENERATED WORK

June 16, 2025 admin Artificial Intelligence, Copyright, IP 0

INTRODUCTION

The rapid advancement in artificial intelligence (AI) is transforming the landscape of the creative industries, giving rise to questions of unprecedented legality and morality. AI-generated paintings and novels as well as symphonies composed in full autonomy compete with human-made works in terms of originality. But such innovations now put to the test the very heart of copyright law: human authorship and originality.

Who owns the rights to the work created by the AI? Should we recognize AI as the author in the current legal terminology? Should credit go to the developer, the user who provided input, or must such content automatically belong in the public domain? These inquiries have fueled debates around the world, and courts, policymakers, and industry leaders are pondering the implications of AI-generated creativity. As jurisdictions across the globe adopt differing approaches to this evolving issue, the legal uncertainty surrounding AI-generated content keeps mounting.

THE AI AUTHORSHIP DILEMMA: A LEGAL CONUNDRUM

Copyright law fundamentally presumes that creative expression originates from humans. Indeed, across the world’s different jurisdictions, it lays down that such creation and its protection should necessarily be ascribed to the exclusive human quality called authorship. This principle makes all legal systems recognize the authors as authors of their literary works, artistic performances, or even music as intellectual property. Thus, the entry of artificial intelligence (AI) into the creative space has thrown such presumptions into a tailspin, unleashing convoluted legal and philosophical debates. The foremost question is: Is AI an author, according to copyright law? If not, who if at all would be entitled to claim ownership of products generated by AI?

The primary obstacle to extending copyright protection to AI-created works is that copyright law shields only human authorship. Most legal systems, including India’s Copyright Act of 1957 unequivocally regard authorship as the qualification of natural persons. An “Author” has been described in Section 2(d) of the Copyright Act of 1957 as the one who creates the work. [1] For literary, dramatic, musical, or artistic works, the author is the individual who expresses original thought and creativity. The act establishes that the first owner of a copyright is the author of the work unless otherwise stated by contract or employment agreements.

Further, Section 13 of the Act states that copyright is granted only to original works. [2] This raises the question that if an AI merely processes and re-arranges existing data and does not create original independent thinking, then can it be termed as original? Since AI does not possess human consciousness and independent judgment, the Courts and Copyright offices, in general, are very hesitant to attribute authorship to machines. This human-centered approach is evident in international case laws, where such AI-generated works are often faced with hurdles in getting copyright protection. Generally, the legal disputes regarding AI authorship are not only confined to India but also to almost every jurisdiction dealing with some variant of the same problem.

In the U.S., the US Copyright Office almost always maintains that copyright protection requires human authorship. This principle was reinforced in Thaler v. Perlmutter, when the court upheld the Office’s refusal to register an image autonomously created by the AI system DABUS; hence, because no human creative input was involved, the work stood to be disqualified from copyright protection.[3]  Further clarification came from the Copyright Office in 2025, stating that authorship is not legally recognized for content that is entirely AI-generated without meaningful human involvement. AI, in various forms, may be involved in the creative procedure; however, only the parts of a work that are the original expression of a human author are protected by copyright. The Office stresses the view that in order for works to be created by authors, human control must be sufficiently higher over at least some of the expressive elements of the work, and whether some creation under the influence of AI is copyrightable shall be considered on a case-by-case basis. [4]

In the U.K., although having a less tolerant and flexible Copyright, Designs, and Patents Act, 1988, it recognizes humans or persons that make the arrangements for the creation of computer-generated work as authors. Such arrangements could make it possible, however unsurprisingly, for AI-assisted works to be qualified for copyright protection. [5]

In China, in the landmark case of Shenzhen Tencent v. Shanghai Yingzun, the court developed a more sophisticated approach in regard to AI-assisted works by granting copyright to a financial article generated by Tencent’s AI software, Dreamwriter, and demanding proof of human intervention in the process. The Shenzhen court considered the idea that the software generated the text but that the creative team met the originality threshold under Chinese copyright law by their involvement in the selection, analysis, and arrangement of financial data. [6]  Similarly, in 2025, the Changshu People’s Court recognized copyright over an AI-generated image where the user used Midjourney and Photoshop, maintaining that prompt engineering and editing of these generated images involved sufficient personal creativity to warrant protection. [7] Whereas, in the Zarya of the Dawn case, the U.S. Copyright Office refused copyright protection for the Midjourney-generated images on the ground that they were not humanly authored and hence could not be copyrighted, but did grant copyright over the textual expression and the selection and arrangement of the materials generated by AI. [8] Such divergent decisions shed light on the evolving debate across the world on the role of human creativity in awarding copyright protection to AI-created works.

While Indian copyright law does not grant express rights to AI-generated works, a landmark decision testing the limits of this question was the copyright application of an AI-generated artwork called Suryast, using an AI tool called Raghav. The Indian Copyright Office had initially conferred copyright registration to the human developer as co-author, putting the test of human involvement in AI-generated works as a possible criterion. Later on, this registration was challenged, hence demonstrating that a grey area in the law continues to exist. [9]

NAVIGATING OWNERSHIP: COMPETING CLAIMS OVER AI CREATIONS

Probably the most intricate and much-debated issue in copyright today is that of defining ownership over AI-generated sources. Copyright laws typically rest on the premise that works of creativity have to be created by human authors. However, now, even machines are able to create art, literature, and music. Laws do not alone accommodate these developments. In the absence of a defined legal recognition of works created by AI, there have been a number of different suggested ownership models, all of which suffer from different legal, ethical, and practical challenges. 

One of these models is to give AI legal personhood, so it might hold copyrights as a human would. However, there is a huge fundamental deficiency as far as law is concerned. The term “author” refers to a natural person according to Section 2(d) of the Copyright Act of 1957,[10] hence an AI cannot be an author in the eyes of law. Moreover, copyright protection under Section 22 of the Act lasts for 60 years after an author’s death, which doesn’t apply to an AI without a lifespan.[11] Besides, the enforcement mechanisms remain vague, AI doesn’t have legal standing to sue or be sued, if the copyright held by AI was to be infringed. Based on these criticisms, most jurisdictions, including India, the USA, and the European Union, have rejected the notion of an independent copyright holder through AI. 

Another proposed model provides for copyright ownership to the developers of the AI system. This is similar to the “work made for hire” theory as applied under Section 17(c) of the Copyright Act of 1957, where the employer is presumed to be the first owner of works by employees when made during the course of employment.[12] If AI were regarded primarily as a kind of nuanced tool and not an independent creator, then its developer would be considered the creator of what the AI produces. This model leaves excellent incentives for innovating AI since the developers would keep their exclusive rights over their outputs. There will, however, be likely instances in which substantial involvement of this AI may be due to heavy prompting or fine-tuning applied by a user. In such cases, the developer’s “ownership” would not hold such ground. 

A further option could be to consider the user to possess copyright of his “computer-generated work”. As Section 2(d)(vi) of the Copyright Act of 1957 states, a computer-generated work is owned by the cause of its creation.[13] However, this opens up a deep inquiry into the kinds and degree of human input needed for the output to meet copyright requirements. The Report of the U.S. Copyright Office of 2025 highlights the centrality of human authorship for copyright to subsist, stressing that works created solely by AI systems without sufficient human creative control are not subject to copyright. Designing this test requires an examination of whether human contribution is significant enough to be deemed an act of authorship-a reflection of the original requirement-thus standing as the foundation of copyright law. A programmatically generated work is considered original if a user beyond mere generic or minimum input works with an AI to produce a novel or a painting. In contrast, an original work generated by AI might no longer be considered original under the law in the Suryast case, where initial copyright grants to an AI-generated painting were challenged. For this, the quality and quantity of human intervention must be examined to ascertain copyright and originality-unless there is a third-party intervention establishing another person as an author.

Another possibility can be joint ownership by AI developers and users. To produce work under the definition of Section 2(z) of the Copyright Act of 1957,[14] there should be many contributory authors whose contributions cannot be separated. If artificial intelligence is to be considered as an advanced tool that will use humans to input, then joint ownership is a viable solution. This master model might be challenged on its practicality with some problems such as disagreement regarding sharing income and licensing with enforcement rights. It is still complex to know how each side’s proportional contributions can be calculated. 

Some experts suggest placing AI-generated works in the public domain, allowing free access without exclusive ownership. This argument rests on the premise that AI is not really creative and only uses already existing datasets. This means that it might promote access and innovation but also deter investment as no profit or revenue is expected. If there is no possibility of reaping profits from AI-generated content, companies are likely to invest less in such kind of advanced AI creative tools. 

CHALLENGES

The emergence of AI-generated content not only brings legal issues but also ethical and commercial problems. Most of the art, music, and literature generated by an AI raise the question of what it means to be made by a person so long as a machine can seemingly mimic such difficult forms of artistic expression. Moreover, the over-acceptance of AI-generated content in cultural and creative industries risks devaluing human creations. Gradually, the structures of incentives that have traditionally inspired artistic innovation might wither away. When creation is freed from associations with human identity and effort, the very age-old purpose of copyright is put in jeopardy: to protect and reward human creativity. These concerns serve as a basis to rethink present frameworks, so that these frameworks will, indeed, work for true creative labor in the age of AI.

Another burning issue surrounds plagiarism and copyright infringement. AI models are trained on datasets, many of which include copyrighted material. Once an AI has generated content that is similar enough to that of some existing work, it would be difficult to say who is liable: the developer of the AI program, the user of that program, or the AI. This legal ambivalence makes prosecution an uphill task while also disrupting the possibility of fair compensation for the original creator. 

AI takes away traditional licensing and revenue models commercially. With potential AI-generated works being the new target of companies, a paradigm change to the intellectual property market will ensue; hence, new frameworks need to be developed with respect to licensing and ownership of these creations and guaranteeing their ethical remuneration in the context of AI usage.

THE WAY FORWARD: AI-SPECIFIC COPYRIGHT LAWS

As AI-generated content begins to outrun established norms for legal regulation, copyright laws need to change from a human-centered view to a more complex view regarding authorship, ownership, and fair use. Laws exist that allow the creativity of humans to be protected, and with the introduction of AI, we need to speak on the differences between AI-assisted and pure AI creations. Here, defining one is considered a fundamental step: whereas in AI-assisted works, humans are heavily engaged in carrying works to completion and copyrighting them, autonomous works can hardly be situated in law. 

In such a scenario, it seems feasible to establish an alternative test, “Significant Human Input,” which averts actors involved in the creation and gauges a meaningful measure of input by human actors. If human creative input is substantiated via creative direction given by a user, modifying and promptly integrating AI outputs in real-time to some extent into a broader creative process, those grant human insights that bestow the right to be personal copyright. This test is in tandem with emerging trends internationally, such as the U.S. Copyright Office decision in Zarya of the Dawn, which only recognized copyright in the human-generated parts of an AI-assisted work. [15]

Also, new licensing models should consider the needs of and try to maintain a good balance between AI developers, users, and original content creators. International cooperation in the matter will also be vital. Setting up global guidelines through the WIPO, for instance, would facilitate a more harmonized mechanism for regulating AI-generated content as well as for the protection of human creativity and innovation.

REFERENCES:

Authored by: Ms. Khushi Pancholi

Fourth Year Student, Institute of Law, Nirma University

 

Be the first to comment

Leave a Reply

Your email address will not be published.


*