Written by Vivek Prasad, Senior Associate at KARM Legal; and Ankita Dhawan, Senior Associate Designate at KARM Legal
If you think copyright is a right that gives the copyright owner the exclusive right to copy the work to which the copyright is owned, then yes, you have got that right!
Copyright protects, among others, literary, artistic and musical works. The creator of a work is its author. The author retains all the economic rights that copyright protection bestows. The author may assign any or all of these economic rights to another entity. The author may also grant licenses for the use of their works by others.
With the advent of sophisticated Generative AI technologies – such as Audiocraft, DALL-E and MidJourney – that allow you to create artistic and musical works, the question that arises is – who is the author of a work produced by Generative AI? The AI system? Or the person who has prompted the AI system to generate a particular work? Or the author of the training data sets on which the AI-generated output is based?
For example, if you prompt an AI system to create a cover of “Yellow” by Coldplay, but in Eminem’s voice, or, to create abstract art in the style of say artist Gerard Richter, who should be entitled to copyright protection for these AI-generated works?
For context, algorithms of AI systems, such as MidJourney, may already be protected. Further, the training data sets would also be protected and would be duly licensed for use by the AI systems (if not licensed, that is a whole new Pandora’s box). Taking the same example, Coldplay’s and Eminem’s music (sound recording, lyrics and composition) and Gerard Richter’s artworks would likely already enjoy copyright protection. However, it gets complicated when we get into the copyrightability of AI-generated musical and artistic works.
In these scenarios, two fundamental questions arise. First, can an AI system be considered the author of an AI-generated work?
AI Systems as Authors
Recently, a District Court in the US, in Thaler v. Perlmutter, was called upon to answer whether a work generated autonomously by a computer can enjoy copyright protection. The Court ruled in the negative, stating that the US copyright law protects only works of human creation. The Court observed that copyright does not protect works generated by new forms of technology operating without any guiding human hand.
But, art, being art, continues to push its own boundaries and evolve. Human inputs in generating art are taking new shapes and forms. This brings us to the second question – if copyright protection is only granted to works with human involvement, what is the minimum threshold of human involvement for an AI-generated work? This question largely remains unanswered.
AI and Human Involvement
A Statement of Policy published by the US Copyright Office states that if a work’s traditional elements of authorship are produced by a machine, the work lacks human authorship, and the Office will not register it. The traditional elements of authorship include literary, artistic, or musical expression or elements of selection, arrangement, etc. The Copyright Office states that if an AI technology receives only a prompt from a human and produces complex works in response, the “traditional elements of authorship” are determined by the technology and not a human.
So, prompts are viewed as instructions given to a commissioned artist. Prompts identify what the prompter wishes to have depicted, but it is the AI system that determines how those instructions are implemented. The copyright law of South Africa defines the author of a computer-generated literary, dramatic, musical or artistic work as the person by whom the arrangements necessary for the creation of the work were undertaken. The position in the UK is also similar.
This implies that if you generate a work using a computer, then you may be the author of such a work. However, works created by generative AI involve a far higher degree of autonomy than mere computer generation. Recently, the Kingdom of Saudi Arabia proposed amendments to its copyright law, which included a dedicated chapter on intellectual property associated with AI.
This states that works created using AI technology will be protectable and that the copyright will be held by the natural person who contributed to its creation. It further states that any creation that is devoid of significant human contribution will not be protectable under copyright. However, what is “significant” remains open.
The UAE has been a regional and global pioneer in recognizing the importance of AI technologies. The UAE copyright law protects smart applications, computer programs and applications, and databases, among others. But, it does not expressly clarify the position on the copyrightability of AI-generated content.
In our opinion, with the advent of AI technologies, while algorithms play a crucial role in the execution phase, the role of human authors at the conception stage remains essential. As AI becomes more pervasive, the human detailing at the conception stage is getting more nuanced, necessitating regulators to find answers to the second question.
This is particularly crucial in the Middle East, where the digital art and NFT sector is rapidly growing. In 2021, the UAE introduced a 10-year National Strategy for the Cultural and Creative Industries, aiming to elevate the GDP contribution of these sectors from 2.6 per cent (as of 2020) to 5 per cent within the next decade! Similarly, one of the Saudi Data and Artificial Intelligence Authority’s strategy objectives is maximizing data and AI’s contribution to realizing the objectives of Vision 2030.
If the digital art economy is truly to be pushed, it would be important for the law to be in step with technology, so that new-age creators are incentivised, and their creations are protected. The future of Generative AI should be based on the principle of “Man and Machine”, rather than “Man versus Machine”.