Author: Yaditya Sandala, Dharmashastra National Law University

Suppose you have prepared a video for your class presentation and have worked extensively to perfectly align the audio and the video, choose the right background music, and explain concepts pertaining to your legal case. You receive a very good score in the presentation. Your presentation is shared on Twitter and your friends upload it on X. Since your video does not have any watermark, someone downloads your presentation, extracts the audio from it, and uses this audio to create an AI-generated video. In this case, who will own the rights to the content?

• The primary legal issue in A.I generated content is OWNERSHIP. The ownership of a particular content is essential because it will protect the person's creative work which was done on his sole hardwork, the court will ensure the person's hardwork is justified and protected.

• Under the Copyright Act of 1957, the challenge in granting rights for copyrighted works in India is the term “author”. The court defines an author as the person who causes the creation of any literary, dramatic, musical, or artistic work, whether computer-generated or not. Ownership of AI-generated content becomes complicated at this point because AI is considered an agent working for a human principal. Courts have often held that a work must be aesthetic and reflect the effort and ideas of its creator. Ownership cannot be easily granted for AI-generated works, as there is no human effort involved, similar to paying a content creator to write a script. Moreover, AI itself is not a human entity capable of holding rights, and a simple text prompt is generally insufficient to claim authorship.

• In the above situation, the court will give you ownership only for the audio generated content because in the case of THALER VS. PELMUTTER the court said that the analysis of the court follows precisely the point you did: copyright protects those portions of a work for which substantial human creative input can be found (selection, arrangement, meaningful modification), and outputs wholly generated by machines without meaningful human authorship are not copyrightable under U.S. law.

• How copyright and A.I are related? It's because A.I will give you the search results of the topic you search is from the internet sources. This will lead to copyright infringement of your work. In the above example you can see that the video which was made for presentation is taken from the already existing audio on the internet from movies or social media reels. This can lead to copyright infringement.

• Now the second issue in A.I generated content and legal issues is copyright because the source which A.I gave you the particular information it was from the internet source. This second issue is relevant in modern time because the use of A.I is being more prominent for many scholars and users.

• There are many case laws, in which India had also raised issues against OPEN A.I for copyright their works which harms their reputation, the licence issues in the case of ANI MEDIA PVT. LTD VS. OPEN AI OPCO LLC in this case, the models have produced false attributions that harmed ANI’s reputation. The suit raises direct copyright infringement and related claims and asks for injunctions and damages.

There are two ways of looking at copyright infringement. On one hand, the copyrighted material could be used as a source or for fair use purposes by an individual. On the other hand, the same material could be used by another to bring innovation by presenting a new interpretation or perspective in view of the material. For example, while writing an article on a particular law, there is always a section known as the Literature Review in which the author has to analyze existing articles or research on the subject. If the author uses AI to provide prompts and gather existing articles or case laws, then this would be in direct violation of copyright laws. Hence, copyright infringement presents both fair and unfair uses of AI-generated content in legal writing. However, in the second perspective, there is a fine line-the courts could reject a claim of copyright infringement if the AI-generated material is used strictly for the literature review section since this is a research and analytical purpose.

Deepfake AI has been used in many videos for content creation in recent times. Even in this context, there are two legal standpoints. Certain YouTubers use deepfake AI to demonstrate how a particular accident or incident occurred. This can fall under fair use since that helps the court imagine what could have happened. But the court does not consider such videos as evidence; they are mere demonstrations for understanding an incident. On the darker side, deepfake AI can also be misused. Recently, there was an incident where a person posted a deepfake video on Twitter showing a leopard entering his neighborhood, which resulted in public panic. Later, the police and forest officials found it to be fake and arrested him. Such misuse of deepfake AI misleads the public and can even mislead the courts to commit errors in delivering justice.

• The Delhi High Court granted an ex-parte injunction restraining the defendants from using the actor's image in deep-fake videos for commercial purposes.

• This establishes how the court protects personality rights (non-copyright) under tort/private rights and recognizes misuse of a person's image, voice, or likeness created by-or altered by-AI/deepfake technology.

• While it is connected with using AI-generated video for someone's identification, the discussion is not really about accident scenarios or court evidence.

The legal implications of AI-generated content in India can be exemplified through the case of Ankit Sahni, an artist-lawyer who developed an artificial intelligence tool named “RAGHAV.” He utilized this tool to produce an artwork titled “SURYAST” and applied for copyright protection, claiming sole authorship over the AI-generated creation. The application was initially rejected by the Copyright Office. Upon appeal, Sahni sought recognition as a co-author alongside the AI system, and the application was temporarily accepted. However, the Copyright Office later issued a notice for withdrawal of the registration. In the existing legal landscape, ownership of AI-assisted works is conferred only upon substantial contribution by a human. These include elements that show creative selection, arrangement, or significant modification of the AI-generated output to establish a threshold for obvious human authorship in the creation.

The determination of authorship and ownership within the interface of copyright and artificial intelligence is challenging. The law still favors human ingenuity as the basis of copyright protection, even if AI can produce original works. Cases such as Thaler v. Perlmutter and Ankit Sahni v. Registrar of Copyrights show how courts expound that ownership rights can only be claimed for creations that represent substantial human input. The creator of the material retains ownership rights in situations where the audio is modified by AI after creation. Without significant human creativity, copyright protection does not cover AI-generated content. With the increased usage of AI tools, deepfakes and the poor exploitation of protected contents require more stringent laws and moral responsibility.

• Thaler v. Perlmutter, 650 F. Supp. 3d 207 (D.D.C. 2023).

• Simply Life India Pvt. Ltd. v. Anil Kapoor, CS(COMM) 2023 (Delhi High Court, Sept. 2023).

• Ankit Sahni v. Registrar of Copyrights, (Copyright Office/Delhi High Court, 2021–2022).

• Narayanan v. Aditya Birla Mgmt. Corp. Pvt. Ltd., (2012) 5 SCC 798 (India).

• R.G. Anand v. Deluxe Films, (1978) 4 SCC 118 (India).

• Star India Pvt. Ltd. v. Piyush Agarwal, (2013) 57 PTC 380 (Del. HC) (India).