The “Infinite Monkey Theorem” by Emile Borel over time became a prophetic insight into the capabilities of developing technology.[1] Artificial Intelligence has currently reached the level where it can create artistic materials by itself with almost no human input.[2] On general understanding, art created by humans is thought of as the manifestation of their thoughts, or in other words a subject of their imagination. Now there are AI tools that can compose music, write poetry, draw paintings and so on.[3] In such a scenario, when AI creates a work of art, the question arises of who actually created it. AI creates art by referring through millions of sources and thus pinning a title of ownership on AI is a matter of discussion. In this article, the issue of copyright in relation to AI-generated content will be discussed. At the time of pioneering AI development through the introduction of platforms such as ChatGPT, this subject becomes all the more important. The article shall go through the subject of how the ownership of AI-generated content is decided, the copyright issues that follow and how such issues can be tackled in the best possible way.
QUESTION OF OWNERSHIP
Artificial neural networks, which are brain-inspired systems that mimic how the human brain works form the basis of artificial intelligence.[4] The works produced by AI can be divided into two categories: “works produced by AI with human intervention” and “works produced entirely by AI.”[5] The creativity in the work can be drawn from the human input in cases where AI creates something with human interference because there is a human input in such cases. In some situations, a human being can be credited as the author. When AI creates a piece of art entirely on its own, without any human involvement, it is unclear who is the true author under the law. For instance, regarding intellectual property, the ChatGPT model is trained on a corpus of original materials, and it is still unknown what the legal precedent may be for reusing this content if it was developed using someone else’s intellectual property.[6]
The photographer claimed he had lost £10k in sales revenue and was unable to pay his lawyer as the issue progressed through court hearings and appeals between 2016 and 2018. The case was whether a monkey could possess the copyright to a photograph.[7] Even though the judge ultimately sided with the photographer since he had organised the photo session and had a key hand in its development, the issue of whether a non-human being can own copyright was exceedingly complicated and hotly contested. At this time, the answer is “no.” Yet, that didn’t stop the case from proceeding to court, taking three years, and nearly driving the photographer into bankruptcy. In the meantime, a Chinese court has decided that an AI-generated work is eligible for copyright protection.[8] It is unfortunate that it is unclear who (if anyone) actually owns the rights to content produced by artificial intelligence.[9]
An AI application that learns from data that has been utilised to train such an AI application may produce the creative works. The data used in this way might have copyright protection and be monetarily valuable. The crucial query is whether copyright infringement occurs when such data are used for machine learning without the owner of the rights having given their consent.[10] The extent of information that the software programme permits the AI to examine will determine the contents, parameters, or other aspects of the invention. AI requires programming in order to provide a result. The work within is based on information that is either publicly available or that is protected by someone else’s copyright since the AI may be able to investigate and analyse the material that is already available. In essence, AI is not capable of producing unique content; instead, the work produced is an adaptation or modified version of already published information. As a result, the identification of AI as a distinct entity and the protection of works independently may result in infringements on the rights of other copyright holders.[11] The ownership of AI-generated work will ultimately rely on the particular circumstances and may need to be decided by a court or other legal authority.[12]
Regarding the authorship question, there are three general options: (1) The copyright system should identify AI as an author; (2) AI-generated works should not have an author and be considered to be in the “public domain”; and (3) sui generis law should be used to defend such works rather than copyright law.[13]
It will be appropriate to make reference to the fair use/dealing concept. It may not be regarded as fair use or trading when the economic value of the copyrighted work utilised for machine learning has been lowered to its owner due to AI-created work. According on the national laws of the various nations, it may qualify as fair use or trading if it does not lower the economic worth of the original work. Typically, copyrighted content used to train algorithms has little impact on its economic value. Hence, the economic worth of the latter is unlikely to change even if a work is produced using an algorithm-powered tool that is entirely different from the copyrighted content utilised for machine learning.[14]
According to India’s Copyright Act, 1957 , a work must be original and have been expressed in a definite form in order to be eligible for copyright protection.[15] The exclusive rights of the owner to perform or authorise the performance of certain acts (such as the reproduction, publication, adaptation, and translation of a work, among others) with respect to a work are referred to as “Copyright” under Section 14 of the Copyright Act of 1957. Furthermore, according to Section 17 of the Act, the creator of the work shall be the original owner of the copyright; but, if the work is produced as part of a contract for consideration and at the employer’s request, the employer is in this case the owner of the work.[16] Originality is not defined in the Indian Copyright Act, and the courts have applied their own case-by-case interpretations of the standard of originality.[17] This could include the use of adequate judgement, skill, and labour; alternatively, the employment of adequate creativity and judgement in the creation of the work. In essence, the project must involve more than just labour and financial investment. An original work immediately acquires copyright when it is made and fixed in a medium. A copyright registration is not a necessity. Additionally, nations that have ratified the Berne Convention for the Protection of Literary and Artistic Works, 1886 (like India), provide simultaneous and automatic copyright protection to any work produced and released in one of their nations. The first recipient of this automatic copyright protection is the work’s author or creator.
Every debate of copyright law must always begin with and focus on authors. In Locke’s labour theory, for example, the author’s claim to the products of his or her intellectual labour is justified when combined with other resources.[18] Although the word “author” is frequently used and mentioned in the Berne Convention, it is not defined there.[19] The Member States of the Berne Conventions had a common understanding of what constitutes an author at the time the Berne Conventions were being debated, thus additional interpretation was not required.[20] Yet, there is a higher possibility of conflict between our intellectual property laws and strategies for creating material without human involvement in the area of patents. In fact, computers are already creating independently in a variety of sectors very useful inventions.[21]
It is clear from Section 51 of the Copyright Act of 1957 that only “persons” are permitted to violate the copyright of a work. Any infringement brought on by AI will be taken seriously because it is still not recognised by the law as a legal entity. It will be far more challenging to assign blame for any infringement brought on by AI in this situation. The issue of granting AI authorship rights may become weak unless a suitable channel and chain can be developed to generate responsibilities for the acts of AI because AI has no independent legal status.[22]
The question today is who might own new works actively made by machines, not whether a photographer can own photographs passively captured by robots. No more than a pen owns the words it writes, neither does a computer programme like a word processor own the text that is entered on it. Yet, AI-powered programmes now create music, news articles, and paintings. Can and ought these acts to be legally protected as they add value? The majority of the time, the response is no. Copyright won’t be applicable until the author can be identified as a human. This policy is frequently cited as encouraging and rewarding innovation.[23] Disputes will therefore likely be resolved case-by-case in light of the facts and circumstances surrounding the creation (e.g., how much of the content required the creativity of the input vs. the generation of the output).[24]
The creation of AI-generated content involves several people. Everyone is involved in this. This necessitates a fresh kind of content licence. Such as, a document that details the extent of everyone’s contribution. Furthermore, we require government agencies that actively control the kind of content produced.[25]
The amount of effort and creativity needed to create an invention using computer-powered linguistic manipulation may be different from what is needed to create other kinds of computer-generated inventions. But the result is the same. Seen from this angle, has society not profited from such diligence given that some organisation would have to invest resources in sorting through a sea of computer-generated claims and figuring out whether a claim is genuinely helpful, if a claim fits all the standards of a patentable invention? In addition to rewarding those who conduct the human work of discovering patentable ideas, allowing such patents would also benefit those who create and enhance the automated invention software itself.[26]
One could draw the conclusion that the Indian legal framework is not ready to address the formation and rights of AI. The inclusion of rights for AI cannot be contested because the Copyright Act of 1957 categorises the author as a person, and until the AI are given a legal standing to this effect. Other strategies include amending the Copyright Act of 1957 to add AI-related works as a separate category or to recognise AI as an author. Yet, neither a change in the law nor the acceptance of AI as a legal entity appear to be likely options in the near future.[27]
AI will continue to play a critical role in intellectual property rights, especially in copyright. The international community has been compelled to consider these questions of authorship and ownership of AI-generated works in copyright law and come up with an agreeable answer for all nations.[28] Although the use of AI to create artwork represents a promising technological leap, there have been legitimate worries that the ambiguous application of copyright law to these creations casts a shadow over their future.[29]
[1] Ben Hattenbach & Joshua Glucoft, Patents in An Era of Infinite Monkeys and Artificial Intelligence, 19 STAN. TECH. L. REV. 32 (2015).
[2] Lexology, Copyright in works created by artificial intelligence: issues and Perspectives,
https://www.lexology.com/library/detail.aspx?g=4513277a-6571-40f1-923d-c09ec5366fdd (last visited 23 Feb 2023).
[3] Supra note 1.
[4] Corrs Chambers Westgarth, Artificial intelligence and copyright: ownership issues in the digital age, https://www.corrs.com.au/insights/artificial-intelligence-and-copyright-ownership-issues-in-the-digital-age (last visited 23 Feb 2023).
[5] Lucy Rana & Meril Mathew Joy, India: Artificial Intelligence And Copyright – The Authorship, Mondaq (22 Feb 2023) https://www.mondaq.com/india/copyright/876800/artificial-intelligence-and-copyright–the-authorship
[6] Joe McKendrick, Who Ultimately Owns Content Generated By ChatGPT And Other AI Platforms?, Forbes (21 Feb 2023) https://www.forbes.com/sites/joemckendrick/2022/12/21/who-ultimately-owns-content-generated-by-chatgpt-and-other-ai-platforms/?sh=74a29ec05423
[7] Susie Kearley, Who Owns the Copyright on AI Generated Content, like ChatGPT?, Medium (22 Feb 2023)
[8] Shayn Fernandez, Who owns the intellectual property rights in AI-generated content?, The Junto Gazette (22 Feb 2023) https://blog.juntolaw.com/who-owns-the-intellectual-property-rights-in-ai-generated-content/
[9] Id.
[10] V. K. Ahuja, Artificial Intelligence And Copyright: Issues And Challenges, ILI. L. R. (2020).
[11] Supra note 5, at 2.
[12] Johan S Daniel, Who owns AI-generated content?, EA Forum (23 Feb 2023) https://forum.effectivealtruism.org/posts/f8jvMin4P3m4Tz5w2/who-owns-ai-generated-content#:~:text=In%20general%2C%20the%20person%20or,have%20a%20claim%20to%20ownership.
[13] Supra note 10.
[14] Supra note 10, at 3.
[15] Indian Copyright Act, 1957.
[16] Supra note 5, at 2.
[17] Supra note 2, at 1.
[18] Ballardini, Rosa Maria, Kan He, and Teemu Roos. AI-generated content: authorship and inventorship in the age of artificial intelligence, Online Distribution of Content in the EU (2019) 117-135.
[19] Jane C. Ginsburg, The Concept of Authorship in Comparative Copyright Law, 52 DePaul. L.Rev. 1063, 1069(2003).
[20] Sam Ricketson & Jane C. Ginsburg, International Copyright and Neighbouring Rights: The Berne Convention and Beyond, 1,OUP, 358 (2005).
[21] Supra note 1, at 1.
[22] Supra note 5, at 2.
[23] Simon Chesterman, AI-Generated Content is Taking over the World. But Who Owns it?, 2 NUSL (2023).
[24] Supra note 8, at 2.
[25] Supra note 12, at 3.
[26] Supra note 1, at 1.
[27] Supra note 5, at 2.
[28] Supra note 10, at 3.
[29] Supra note 10, at 3.
Auhtor: Ronsha Roys Anna from NUALS, Kochi
Juris Centre 