ACIPR Logo

Alliance Center for
Intellectual Property Rights


WHO OWNS AI-CREATED ART? NAVIGATING COPYRIGHT LAWS IN THE AGE OF GENERATIVE AI

September 15, 2025

*Ms. Khushi Verma

INTRODUCTION:

Artificial Intelligence (hereinafter “AI”) has revolutionized how we interact with technology, create content and solve problems. It comprises a multitude of different technologies like LLMs, computer visions, etc. AI basically enables such technologies to replicate human intelligence by feeding it vast amounts of data and information within milliseconds. Among its most transformative applications is generative AI (hereinafter “GenAI”) which is a technology that can create new text, images, audio, and video with minimal human input. Its rise is reshaping entire industries has raised important legal and ethical questions, especially around copyright and intellectual property.

GenAI refers to a class of AI that can create new content. Unlike earlier versions of AI, which simply identified patterns in existing data, GenAI produces new works based on the input it receives. This technology includes popular tools such as ChatGPT, DALL-E, and various music and video generators, all capable of crafting unique outputs in seconds.

Generative AI is not just about automating tasks; it is about amplifying creativity and productivity. However, as these systems continues to evolve, it disrupts long-standing ideas about human creativity, pushing us to restructure how copyright law responds to the same.

ARE AI-GENERATED WORKS "ORIGINAL" ENOUGH FOR COPYRIGHT PROTECTION?

Copyright law is built on the idea of protecting "original" works. But are creations made by AI original? GenAI work is rather seen as an adaption or modification of existing information that it has been trained on. But some theories also suggest that AI can learn from the datasets and create novel work and the very prompt used for generating the work is also a creative expression. Over the years there have been several tests to judge whether a certain work should be protected. The most popular of these tests are the Sweat of the Brow test, the Modicum of Creativity Test, and the Skill and Judgment Test.

The Sweat of the Brow test was originally discussed in the case of University London Press v. University Tutorial Press where the Court decided that time and effort spent were enough for copyright protection. This test looks at the effort, skill, and labour invested in creating a work. However, since AI can produce content in seconds without human effort, works made by AI often struggle to meet this bar.

The Modicum of Creativity test says that only works with a minimum level of creativity qualify for copyright protection. AI-driven content can be creative, but if it just repurposes existing data, does it really pass this test?

A fair middle ground between these two tests is the Skill and Judgment Test. This is the approach that is generally undertaken by Indian courts to determine the copyrightability of a work. It requires an author to apply his skill and judgment in creating work and the work should have a minimal level of creativity. AI-generated works do not always meet traditional standards of originality. The law is still undecided on when and how such works should be protected. With the threshold for originality often set low, some AI content may end up protected, but legislative gaps remain.

THE OWNERSHIP DILEMMA:

Once a work is deemed original, the next big issue is ownership. Traditionally, the person who creates the expression of an idea owns the copyright. Someone who just has the idea itself cannot deem it to be copyrighted. When AI comes into the picture, it is very difficult to pinpoint who is the creator as AI is only outsourcing the expression of our ideas to execute it. Because of this issue, we need to delve into AI and whether AI can be considered as a separate entity from its creator and thus, can be deemed as the creator of the work. Or is there a possibility of co-authorship between the AI and the end user?

Since AI does not have legal personality yet, it cannot be called as an independent author. The focus then shifts to who “caused” the work to be created. Then can the end user be termed as the author? Section 2(d)(vi) of the Copyright Act, 1957 states that in the case of computer-generated work, author would mean any person “who causes work to be created.”

In USA, it has been clarified that works created solely through generative AI is not copyrightable. However, it does hint at the possibility of the end user owning the copyright if the prompt is sufficiently detailed such that it shapes the ultimate work. This is in line with the Indian jurisprudence on copyrights. Following the Skills and Judgment Test, the Indian courts are of the opinion that for a work to be protected, there has to be a certain degree of creativity and skill on the part of the author. So, whoever carefully works the prompts in such a manner so as to get a very specific result may be the copyright.

FAIR USE AND COPYRIGHT INFRINGEMENT IN AI TRAINING AND OUTPUTS

A significant concern arises in determining where the line between fair use and copyright infringement lies, particularly since AI models are trained on massive datasets that often include copyrighted materials.

At first glance, it would seem like this unauthorised use of work is a misappropriation of copyright, but the situation requires some nuance. The fair use doctrine is a widely recognized exception in cases of copyright infringement. The general idea is that as long as the copyrighted material is ‘transformative’ in nature, there exists no infringement. In the Authors Guild, Inc, v. Google, Inc., Google's digitization of books to create a searchable database was held to be transformative and thus fair use.

THE FUTURE PATH:

Generative AI is being adopted in so many sectors like education, healthcare, business, etc. and it will continue to transform how industries work. Moreover, it is rapidly advancing. It is becoming better at handling data and using it to furnish further content. Each version can be said to be better than the previous version. This is all the more reason to find proper ways to deal with the complexity of generative AI and copyright infringement.

The best way to deal with the same is creating a legal framework which addresses the copyright law questions raised by GenAI. In India, an eight-member expert panel was constituted by the Ministry of Commerce & Industry to adequately address the unique challenges presented by GenAI. While the official report is not out yet, the panel is said to be thinking of bringing a licensing scheme to balance the interests of creators and AI companies.

Moreover, the panel is also attempting to draw a distinction being “AI-assisted” work and “AI-generated work”. This might play out to be crucial in resolving the ownership dilemma.

Since the data used to train GenAI is not limited to just one country and is sourced globally, simply having domestic regulations may turn out to be futile as differing laws of different countries may still lead to uncertainty. Therefore, international harmony is just as essential as domestic laws. This will help in addressing the legal and ethical challenged being faced globally due to GenAI.

CONCLUSION

Generative AI has been one of the biggest technological advancements so far. The advancements that have been made in creativity and productivity is unparalleled. However, we cannot deny that the same development also poses certain risks and challenges pertaining to the basic principles of copyright law.

Laws need to evolve to keep up with the progress taking place in the digital world which is highly unregulated. Developing effective laws will require collaboration between AI developers, policymakers, and authors of copyrighted work to reach holistic solutions that addresses these issues adequately.

REFRENCES:

  1. CONGRESS.GOV, https://www.congress.gov/crs-product/LSB10922 (last visited May 12, 2025).
  2. BRITANNICA, https://www.britannica.com/technology/artificial-intelligence/Methods-and-goals-in-AI (last visited May 12, 2025).
  3. Maheshwari & Co., Generative AI & Copyright Law In India: Who Owns Machine-Made Works, MONDAQ (May 12, 2025, 9:35 PM), https://www.mondaq.com/india/patent/1653344/generative-ai-copyright-law-in-india-who-owns-machine-made-works
  4. GEEKS FOR GEEKS, https://www.geeksforgeeks.org/artificial-intelligence/artificial-intelligence-101/ (last visited May 12, 2025).
  5. R. Singh, Understanding the Concept of Originality Under Copyright Law in India, LAW MANTRA (May 12, 2015, 9:30 PM), https://journal.lawmantra.co.in/wp-content/uploads/2015/08/11.pdf
  6. University London Press v. University Tutorial Press, Ltd., [1916] 2 Ch, 601 (Eng).
  7. Feist Publications, Inc. v. Rural Television Service Co., 499 U.S. 340 (1991).
  8. R.G. Anand v. Delux Films, A.I.R. 1978 S.C. 1613 (India).
  9. The Copyright Act, 1957, § 2(d)(vi), No. 14, Acts of Parliament, 1957 (India).
  10. GOV INFO, https://www.govinfo.gov/content/pkg/FR-2023-03-16/pdf/2023-05321.pdf (last visited May 13, 2025).
  11. Authors Guild v. Google, Inc., 804 F.3d 202 (2d Cir. 2015).

Author:
* Ms. Khushi Verma
Student at ILS Law College, Pune.

Disclaimer: The opinions expressed in the article are the personal opinions of the author. The facts and opinions appearing in the article do not reflect the views of the Alliance Centre for Intellectual Property Rights (ACIPR) and the Centre does not assume any responsibility or liability for the same.