top of page

Artificial Intelligence and the Copyright Quandary: A Legal Discourse on AI-Generated Artwork



Written by Syed Suhaib and Sankalp Mirani,


The integration of AI in creative processes challenges traditional notions of authorship and ownership. We need to rethink copyright law to account for the collaborative nature of AI-generated works and ensure fair distribution of rights and rewards. - ~ Urs Gasser (Executive Director of the Berkman Klein Centre for Internet & Society at Harvard University).


Introduction

Every artistic composition harbours the latent potential to metamorphose into a magnum opus, encapsulating the artist's or creator's profound wisdom as they breathe life into their creation through profound intellect and unbridled creativity. This correlates with one of the fundamental principles of Intellectual Property Rights, which resides in the preservation of the creativity brought forth by artists, thereby staunchly upholding the sanctity of their creations. In the general precept of Intellectual Property Rights, the conventional notion is that creations are attributed to human ingenuity. Even though machines can assist in enhancing creativity, the foremost merit of copyright is attributed to the human source of that creativity. The preeminent music, image, and art sectors encountered a quandary when Artificial Intelligence (AI) technology was harnessed to create original artwork. The AI models that were utilised to achieve this apparent ‘original artwork’, were trained using specific existing data sets and consequently developed the ability to craft remarkable art with absolute precision and creativity.


Nonetheless, apprehensions surfaced when the creators of the original artworks, whose ‘initial creations’ were employed for the training of the AI models, raised queries regarding the legality of the unapproved utilization of their artistic endeavours. In India, the Indian Copyright laws regulate the original artistic and creative works which are incorporated under the Copyright Act of 1957. Section 2(d)(vi) of the Copyright Act 1957 defines an ‘author’ and provides an insight into the authorship of ‘computer-generated work’. For any literary, dramatic, musical or artistic work which is computer-generated, the person who causes the work to be created will be considered as the author. To abridge, the copyright laws in India are yet to recognise AI-generated artwork. In the absence of any statutory regulations, the conundrum that was brought forth was, whether the utilization of pre-existing data (original artworks) in training AI models constituted an infringement upon the intellectual property rights of the original creators. Consequently, this crossroad has led to this apparent quandary between copyright and artificial intelligence. Pertinently, this junction requires not just a discourse but rather a profound statutory framework backed with implementations to safeguard the principles of intellectual property rights. 


AI Art Generation and Legal Annotations on Infringement

The technology of Artificial Intelligence as we are witnessing today, has not always been like this, otherwise, the inclusion of machine assistance might have been inculcated in the copyright law till now. However, back in 2012, the state-of-the-art models were able to generate only a benign amount of artwork out of the thousands provided. Consequently, the incumbent sterility of AI can generate that outcome within seconds. The legality of the output generated by AI was questioned.  The dominance of AI and its eccentric nature was brought into question. For instance, should one desire to hear The Weekend’s "Save Your Tears" sung in the style of J Cole, there are AI-driven technologies with the capacity to execute such a feat. These technologies encompass Jukedeck, Magenta by Google, Amadeus Code, MuseNet, etc. In the United States, a suit was filed by the artists Kelly McKernan, Karla Ortiz and Sarah Anderson against Stable Diffusion (created by Stability AI), Midjourney and Dream (who are the most mainstream AI art generators) for utilising a multitude of images without procuring the requisite consent from the original artists. Another instance was that Getty Images also filed a lawsuit against Stability AI for the same purpose. In this context, Anand Nishanth, a legal expert specializing in Intellectual Property (IP), asserted that copyright can solely pertain to concrete creative works rather than a particular style.


Furthermore, he posited that in the aforementioned legal actions, the contention centres on the allegation that their original works are being replicated to produce additional identical images. This serves as an alternative argument in the pursuit of holding AI-powered art generation companies accountable for potential infringements of Intellectual Property Rights. The art enterprises that employ Artificial Intelligence technologies to produce artistic creations from pre-existing data should be mandated to divulge the sources of data employed in the training of their AI models and provide equitable compensation to the artists whose contributions facilitated the development of said AI models. This approach allows AI-powered art establishments to sustain the creation of novel works without trespassing upon the intellectual property rights of other creators, signifying a momentous stride in the regulation of the limitless potential inherent in Artificial Intelligence technology.

 

Revisiting Copyright Laws: The Imperative for Reform

The copyright acquisition process varies by jurisdiction, yet a common stipulation lies in the necessity of originality. Within the Indian legal framework, section 13 of the Copyright Act 1957 explicitly mandates that to qualify for copyright protection, a work must exhibit inherent originality[TJ1] . This consequently raises an inquiry pertaining to the classification of artwork produced through AI-powered technology as being truly original. The term ‘original’ is not defined in the Copyright Act of 1957, however, the Hon’ble Supreme Court in Eastern Book Company & Ors. v. D. B. Modak and Anr (2008) 1 SCC has interpreted the term ‘originality’ and what creation can be constituted as an ‘original creation’ under law. To determine the originality of a work, two assessments are considered imperative to conduct; 'Modicum of Creativity’ and ‘Sweat of the Brow’. The ‘Sweat of the Brow’ metric, is a broad criterion that bestows copyright protection upon those who invest labour and resources in their creations.


The 'Modicum of Creativity’ is a creative standard that demands an invention or production to possess uniqueness and non-obviousness. By calibrating the intersection between the two assessments, the term ‘original creation’ can be interpreted. The United States Supreme Court also denied copyright protection to an AI-generated image based on lack of originality. Stephen Thale’s application before the U.S. Supreme Court for his scientific innovation was also denied by both the U.S. Copyright Office and a district court presided by District Court Judge Beryl A. Howell. This substantiates that the comprehension of the human element in AI holds significant sway in conferring protection, thereby necessitating a critical reassessment of the existing legal framework. AI-powered technologies possess the capacity to either craft original work or substantially contribute to its innovation. These AI-powered technologies can assimilate pre-existing data inputs and generate entirely novel creations, which in themselves can meet the criteria for the 'Modicum of Creativity.'


Two questions are being posed here; First, Should AI-developing companies be allowed to use the original data to train their AI? Second, how can the use of an artist's original creation be justified to generate other identical or non-identical works? Although these questions do not yield definitive answers, an attempt has been made to address these queries. In response to the first inquiry, AI, even in its preliminary and experimental stages, has proven to be a valuable collaborator, albeit one that necessitates access to data to fulfil its role effectively. As for the second question, the law should govern equitable AI usage and establish mechanisms for compensation to the original artists for their work. AI enterprises should also be mandated to provide comprehensive disclosures pertaining to the training of their AI models to the relevant authorities. This oversight ensures the utilization of safe and appropriate data in the training of AI models.

 

Conclusion

The integration of AI and its quandary with not just copyright laws but rather within the sphere of Intellectual Property Laws has posed intricate barriers to the existing framework. The very essence of 'originality' is under scrutiny, as AI demonstrates its ability to craft unique and innovative works, often drawing from existing data sources. The legal landscape, both in India and the rest of the world, underscores the paramount significance of human involvement in conferring copyright protection by performing a critical review of the existing legal framework. Another quandary that looms large revolves around the permissibility of AI companies utilizing original data for their AI training and the justifiability of employing an artist's initial creation as a basis for generating either identical or distinct works which call for parallel developments to ensure a check on the powers and utilisation of Artificial Intelligence in the hands of bad actors. The copyright conundrum of AI-generated artwork is a testament to the evolving integration of technology and human creativity which can result in flamboyant outcomes which may, at the same time, be equally dangerous in its legal scrutiny.


The recent news surrounding the AI quandary in the record label industry has just piled the AI concerns with claims that it can take existing practices out of business. However, with empirical evidence, it can be reasoned that AI is not as advanced as off right now, although as I finish writing this sentence, it may have become more advanced by then, such is the pace of advancement of AI. Even with this pace, the AI advancement seems to lack certain intricacies which is much evident now that, a particular thing may be AI-generated. Nevertheless, it is still incumbent for the legal framework to adapt and safeguard the rights of the original creators and to draw a distinction between machine and human creativity and the result of their integration. Consequently, the need for reform becomes evident and the introduction of the AI Act by the European Union marks a substantial step towards underscoring the necessity for legal frameworks that can accommodate the integration of AI while upholding the principles of creativity, originality, and artistic creation.  


230 views0 comments

Comments


bottom of page