GENERATIVE-AI
In the contemporary world, Artificial-Intelligence(AI) is taking over almost all the crucial industries in the world market. In late November of 2022, ChatGPT was released, a product developed by OpenAI, a form of Generative-AI which took the world by storm. There are several other similar products such as Claude, Meta AI, Microsoft Copilot etc. Deep learning models that can produce top-notch text, graphics, and other content from the training data are referred to as generative AI.[1] Due to the rise in the usage of Generative-AI by corporates as well as private individuals, there is a sudden surge in legal disputes related to the ownership of such AI generated content and whether such content can be copyrighted. Another emerging question is whether any content generated, on the basis of any data on which the AI was trained is copyrighted, would lead to Copyright infringement.
OWNERSHIP AND COPYRIGHT OF AI GENERATED CONTENT
The ownership of the content is not an issue in case the AI is used for personal use, and will not lead to any infringement of any copyright. The legal term “copyright” refers to the ownership rights that authors and artists have over their creative works. Books, music, paintings, sculptures, films, computer programs, databases, ads, maps, and technical drawings are among the works that fall under the purview of copyright protection.[2]
In India, Copyright Act, 1957 governs matters related to copyright. Section 52[3] of the Act states about the doctrine of fair use and scenarios in which certain acts or contents cannot be considered as copyright infringement such as acts done or content created for personal or private use. In the case of Civic Chandra & Others v. C. Ammi Amma & Others.[4] the court gave tests to determine whether there has been infringement of Copyright or not, namely; the quantum and value of the matter taken in relation to the comments or criticism; the purpose for which it is taken; and the likelihood of competition between the two works.”
The doctrine of Fair use was established in USA in the case of Folsom vs Marsh,[5] wherein four factors were set forth to determine a work to be fair use; “The nature and objects of the selections made; The nature of the original work; The amount is taken; and the degree in which the use may prejudice the sale, or diminish the profits, or supersede the objects, of the original work.” Later on this was codified under the Copyright Act, 1976 of the USA. It can be inferred that use of AI for personal and private use will neither lead to any dispute as to ownership nor lead to copyright infringement.
The real issue arises when the content is intended to be used for some further commercial purpose, where determining the ownership of such content becomes essential. One may think it is OpenAI in the case of ChatGPT, while others may think it is the person who uses the product by providing a prompt; or the person whose content has been used as data to train the AI such as ChatGPT and other similar products.
In India, The Copyright Act, 1957 provides protection to the work of an author which was originally authored by him, such works can include literary, dramatical musical and artistic works from unauthorised uses. Copyright protects the expressions and not the ideas. There is no copyright in an idea.[6] Section 2(d)(vi) of Indian Copyright Act, 1957[7] provides that the author of a computer generated content will be the person who causes such content to be created, but it does not talk about Generative-AI content. In India copyright subsists only in (a) original literary, dramatic, musical and artistic works; (b) cinematograph films; and (c) sound recording.[8]
In USA Copyright Act, 1976 governs the aspects related to copyright, Congress included a list of eight works of authorship in the Act itself i.e. literary works; musical works, including any accompanying words; dramatic works, including any accompanying music; pantomimes and choreographic works; pictorial, graphic, and sculptural works; motion pictures and other audiovisual works; sound recordings; and architectural works.[9]
AI AS OWNER
AI alone, without any human input, cannot generate any content. The generated content is wholly based on the prompt given by the user according to his requirements. They lack independen thinking as that of a human. Human authorship is an essential condition for copyright protection and ownership. On a combined reading of Section 2(d)[10] and Section 17[11] of the Copyright Act, 1957 it can be understood that copyright can be claimed only by a natural person. In the case of Navigators Logistics Ltd. v. Kashif Qureshi & Ors.[12] the court while deciding a case related to stealing of confidential copyrighted information put forth its view that, the author of a literary work must be a natural person and not an artificial person, and an author must create a piece of literary work using his skill and judgment to receive protection under the Copyright Act, 1957. Copyright cannot exist if there is neither skill nor judgment used in a compilation.
A similar scenario exists in the USA, an AI generated work can only be copyrighted only if the content had significant creative input from a human author, the content is not a derivative of someone else’s work, and it was not created by AI alone.No application can be filed for copyright of a picture, in the name of an animal, that was clicked by itself.[13]
Suryast
Suryast was an art piece generated by an AI tool named RAGHAV( Robust Artificially Intelligent Graphics and Art Visualizer) developed by Mr. Ankit Sahni. The United States Copyright Office refused copyright registration of Suryast in the name of RAGHAV, and despite the arguments made by Sahni, the registration was refused. The Indian Copyright office initially granted registration of Suryast with Mr. Sahni as the co-author of the art generated, but subsequently it was also withdrawn making the status of Indian Legislation unclear in this aspect.
The need for the human element was also emphasized in Thaler v. Perlmutter[14] where it was held that there can be no ownership of work generated by AI tools developed by a person on his own. Though it is visible that AI cannot own or have a copyright for the content generated by it, in its own name, it is insufficient to conclude such a thing till any proper guidance notes or notifications are released by the appropriate authorities in such matters.
PROMPT GIVER AS OWNER
The person giving the prompt can in some cases be regarded as the owner of the content genrated by the AI and can be granted a copyright for such work, but there are certain factors that influence it.
The originality of the content is one of the most important factors that influence its ownership[15]. The content generated if in any way is identical or similar to the original work from which such output is generated will not be entitled for a copyright. Any efforts made to differentiate it from the original should be considered while deciding such matters, as was stated in Eastern Book Company v. D.B. Modak[16], where the court held that the derivative work needed to incorporate individual skill, labour, capital, and a minimum amount of imagination in order to be more than merely a copy of the original work. This case was related to a journal known as SCC which published Supreme Court cases with several additions.
This case paves the way for another aspect i.e. the efforts of the prompt giver. Simply doing the work of putting a one liner prompt will not be considered as a significant effort by the prompt giver. Although there is no proper criteria for determining the significance of the prompt giver’s effort, there should be reasonable efforts put by him for the generated content to be copyrighted. Further guidelines were released by the US Copyright Office, stating that “a work containing AI-generated material will also contain sufficient human authorship to support a copyright claim,” in the event that an author had creatively arranged or significantly altered AI-generated works. The human-authored portions of the work, on the other hand, are “independent of” and “do not affect” the copyright status of the AI-generated content itself, therefore in these situations, copyright will solely protect those portions of the work.[17]
Zarya of the Dawn
Zarya of the Dawn was a graphic novel which was made by selecting and arranging the images and texts generated by a Generative-AI platform called midjourney. The novel was proposed to be copyrighted. The United States Copyright Office denied registration for the images, but allowed registration of the selection and arrangement of the images and text generated by the AI. It was shown by him that such arrangement and selection were completely his own efforts and with interference. In the same guidance statement, the US Copyright Office also reaffirmed that protection remains for underlying original works, even where they’re enhanced by technological tools.[18]
OTHER CHALLENGES
Under the Indian Copyright Act, 1957, Chapter V[19] of the Act talks about the term for which a copyright can be granted, stating that the maximum time limit shall be sixty years after the death of the Author, be it anonymus or pseudoanonymus work, posthumous work etc. Whereas under US Copyright Act, 1976[20] The Copyright exists for a period of seventy years after the author’s death in general cases, and in case of anonymus or pseudoanonymus work the copyright exists for a period of ninety five years from publication or one hundred and twenty years from date of creation, whichever expires earlier.
If in case a copyright is issued in favor of any AI or AI developer which has perpetual existence, the time period for which a copyright exists will be infinite. This will defeat the whole purpose of the statute.
Another challenge relates to the liability for such AI generated content. When any AI generated content causes any injury to any person, it is difficult to ascertain who will be held liable for such injury, now the person claiming the ownership of such content be it the AI, the developer or the prompt giver, will try to deny his liability over such content.
\
WAY FORWARD\
Shri. Som Parkash, Union Minister of State for Commerce and Industry, in a written reply in the Rajya Sabha, made a statement that “Therefore, while Artificial Intelligence (AI) and related innovations is an evolving stream of technology, the current legal framework under the Patent and Copyright Act is well-equipped to protect Artificial Intelligence generated works and related innovations. Presently, there is no proposal to create any separate right so ram end the law in the context of AI-generated content.”[21] But this statement is valid only upto some extent. There are many aspects which are left unrecognized by the Indian Intelectual Property Law in relation to AI.
In its guidelines, the US Copyright Office stated that it will keep an eye on this new technology and perhaps provide more instructions in the future. Because of this, the existing guidelines on “sufficient human authorship” may eventually be in favour of copyright protection being granted for works created by AI, or for parts or variations of them.[22]
CONCLUSION
The use of AI is growing at a significant level even in India. The current framework for regulation of AI and copyright is insufficient to tackle the disputes that may arise in the future due to the growth of AI. It needs to be upgraded in a way taking into consideration all the related aspects by brainstorming about the various scenarios that may arise in the future, while keeping in mind the current status.
[1] Kim Martineau, What is generative AI?(20 April, 2023), https://research.ibm.com/blog/what-is-generative-AI
[2] Copyright, World Intellectual Property Organisation, https://www.wipo.int/copyright/en/
[3] Copyright Act, 1957 ; § 52
[4] (1996) 1 KLJ 454
[5] 9. F.Cas. 342 (C.C.D. Mass. 1841)
[6] A HAND BOOK OF COPYRIGHT LAW, Government of India,Department For Promotion of Industry and Internal Trade, Ministry of Commerce and Industry, .https://copyright.gov.in/documents/handbook.html#:~:text=The%20Copyright%20Act%2C%201957%20protects,expressions%20and%20not%20the%20ideas.
[7] Copyright Act, 1957 ; § 2(d)(iv)
[8] Copyright Act, 1957 ; § 17
[9] COPYRIGHT LAW IN THE UNITED STATES, BITLAW, https://www.bitlaw.com/copyright/obtaining.html
[10] Copyright Act, 1957 ; § 2(d)
[11] Supra note 9
[12] (2018) 254 DLT 307
[13] Naruto vs Slater, 888 F.3d 418 (9th Cir. 2018)
[14] No. CV 22-1564 (BAH) (D.D.C. Aug. 18, 2023)
[15] Harshal Chhabra* and Kanishk Gaurav Pandey, Balancing Indian Copyright Law with AI-Generated Content: The ‘Significant Human Input’ Approach, ndian Journal of Law and Technology( Feb 26, 2024), https://www.ijlt.in/post/balancing-indian-copyright-law-with-ai-generated-content-the-significant-human-input-approach#:~:text=It%20was%20held%20that%20human,can%20claim%20copyright%20or%20not.
[16] (2008) 1 SCC 1
[17] Copyright Ownership of Generative AI Outputs Varies Around the World, Cooley (Jan 29, 2024), https://www.cooley.com/news/insight/2024/2024-01-29-copyright-ownership-of-generative-ai-outputs-varies-around-the-world
[18] Supra note 17
[19] Copyright Act, 1957 ; Chapter V
[20] Copyright Act, 1976 ; § 302
[21] Ministry of Commerce & Industry, Existing IPR regime well-equipped to protect AI generated works, no need to create separate category of rights, Press Information Bureau, Delhi (FEB 9th, 2024 8:55 PM), https://pib.gov.in/PressReleasePage.aspx?PRID=2004715#:~:text=Therefore%2C%20while%20Artificial%20Intelligence%20(AI,generated%20works%20and%20related%20innovations.
[22] Copyright Ownership of Generative AI Outputs Varies Around the World, Cooley (Jan 29, 2024), https://www.cooley.com/news/insight/2024/2024-01-29-copyright-ownership-of-generative-ai-outputs-varies-around-the-world.
Author: Sisir Rao
Juris Centre 