– Sheeba Masoodi
Introduction
If art is an expression emanating from human emotion, can AI be an artist? AI generated art has raised a handful of concerns germane to Intellectual Property Rights, accountability and governance in particular. The emergence of AI image generators has sparked unease in the art world as these generators are trained on datasets of images, which are retrieved from open-access sources on the internet. This has led to possibilities of AI using copyrighted works without authorization. However, even though AI generates the artworks, human involvement remains inevitable. AI being a diffuse term that corresponds to a web of human actors and computational processes interacting in complex ways, may lead to obfuscation of accountability of human actors.
The copyright laws in India are governed by The Copyright Act of 1957, is limited in its scope in the realm of AI, however through the 1994 Amendment it granted authorship rights of an artwork generated by a computer to the person who affected such creation of the artwork. This has led to a myriad of scenarios concerning infringement of copyrights unregulated. The current copyright laws in India do not address the questions as to the liability of engineer who uses copyrighted works to train an AI program to generate works of art. This question becomes of importance as AI art has begun to gain recognition, as made evident by the case of “Edmond de Belamy, from La Famille de Belamy” which is an AI-produced artwork, sold for $432,500. This article traces through the inadequacies of current copyright laws, shedding light on their inability to address emerging challenges in the age of artificial intelligence.
Copyrights to AI Generated Artwork – A Sound Approach?
In the past decade, AI has garnered significant attention from both those whose rights it threatens and those tasked with safeguarding against such adversities. The debate surrounding AI and copyrights has remained anchored to elements of authorship, creativity and originality. When assessing the possibilities of as to who can be assigned the authorship of an AI generated work, several stakeholders have been identified including firstly, the AI program itself, secondly the programmer responsible for the development of the technical program and thirdly, the human involved in creating requisite arrangements of the creation of such artworks by the AI. Given the variety of stakeholders, it becomes crucial to analyse the concept of ‘authorship’ and determine which approach to granting authorship would be the most appropriate.
Attributing authorship to AI would be legally untenable, as authorship rights are inseverable from transfer, ownership, licensing and moral rights as such rights cannot be utilised by an AI program. As far as the second possibility is concerned, declaring the programmer as the author would create a legal anomaly, as it would impose liability on them for every outcome produced by the program, even in instances where they had no involvement in the eventual creation of the artwork.
Till present, copyright laws have been human-centric, recognising only humans as authors for their sole benefit. It has been a well-established principle that copyright laws were aimed at protection of human creativity and originality to promote innovations, until the advent of AI. This was further backed by the “Test of Creativity” commonly employed to determine whether a copyright arises out of a particular piece of work. According to this approach, if an artistic work satisfies a minimum level of creativity, it qualifies for legal protection. However, assuming that an AI generated artwork passes the Test of Creativity, the rights accruing to such authorship remain unregulated. The Indian Copyright Act (1957), after the 1994 amendment has introduced the concept of “computer generated artworks”. It recognises in the “person who causes the work to be created” as the author of a computer-generated artwork.
In Delhi High Court in Camlin Pvt. Ltd. v. National Pencil Industries, the court has held that authorship requires the involvement of a natural person. Hence, in India thus far the scope of authorship has not been extended to non-natural entities.
Further Section 22 stipulates that a copyright shall subsist in an artwork published within the lifetime of the author, and up to 60 years after his death pointing that authorship is limited to natural persons. Moreover, Section 51 of the Act stipulates that an infringement of a copyright can be done by a “person”.
This creates a legal vacuum in scenarios where the person causing the creation of an artwork becomes untraceable, allowing a space to escape liability for copyright infringement. Given that the technical program itself has not been given legal recognition as an author under The Copyright Act (1957), the possibilities of imposing liability are thinned further.
Further, the process underlying generation of artworks by AI programs renders it unqualifiable for copyrights, since an input and datasets available in the public domain which it blends to create an output are a pre-requisite for such generation. This raises questions of originality of AI generated artworks, and the validity of authorship rights arising thereof. For a piece of art to be deemed original, it must be an origination from the author’s own creative efforts and not a mere duplicate recreation. Subjective theory emphasizes the individual author’s subjective creativity and personal expression as the basis for originality. This approach would exclude AI produced artworks from copyright protection on the grounds of originality. Ankit Sahni’s artwork developed by an AI tool named RAGHAV (‘Robust Artificially Intelligent Graphics and Art Visualizer’). Ankit Sahni was able to produce an outcome from RAGHAV by feeding a photograph and an image of “The Starry Night” by Van Gogh as an input. The work so produced was a blend of the two images, and a derivative in contrast to an original creative work, as stated by Review Board of the US Copyright Office while denying registration to the artwork.
It is noteworthy that Ankit Sahni and RAGHAV were registered as co-authors of the artwork, although the legal framework does not make a provision for granting authorship to an AI tool.
The Indian Copyright Office seems to be at a predicament in dealing with applications seeking registration of artworks made by AI. With the lack of precedent as well as an adequate legal framework, the grounds of registration of Ankit Sahni’s work remain vague. Moreover, such a practice could serve as a precedent for future applications which may push the Copyright Office into an impasse
The Next Rembrandt serves as another fine example of limitations of originality of AI. The said painting was created after feeding the program with data on existing works of the renowned artist Rembrandt.
Granting copyrights to AI-generated artworks could significantly undermine artists’ livelihoods. The ease and speed of creating art using AI tools could flood the market, leading to potential inflation in the art industry. With AI’s accessibility, anyone can produce art, enabling businesses to mass-produce and commercialize such works. This not only diminishes the exclusivity and value of traditional art but also poses a threat to its very essence by prioritizing quantity over creativity and authenticity.
Allowing a person who merely prompts an AI program to generate an artwork, to share the same pedestal as an artist who puts in efforts, creativity and diligence into developing an idea and giving it the form of a piece of art, would endanger principles of the historically recognised Sweat and Brow theory. This would allow a copyright to be attributed to a person who did not creatively contribute to an artwork, defeating the objective of the copyright system.
Conclusion
The advent of AI has created lacunae in the Indian copyright law. With minimal legal developments taking place to address the newly arising concerns, a state of confusion prevails as conflicting interests and uncertainties surround the rapid increase in the usage of AI. While evolving notions of authorship globally step towards bringing AI within the garb of copyright law, India is yet to establish its position. According authorship to AI seems to be an unsound choice, as it could risk diluting human creativity.
So far, the only formal step taken in the face of technological growth is the Parliamentary Standing Committee Recommendation on ‘revisiting of IPR legislations and implementing a strong IPR framework’. Being at a preliminary stage in this sphere, India is served with an opportunity to come up with a deliberately crafted framework to avoid flip-flops in the future. Though present laws could be interpreted with more flexibility to tackle the issues, however, a far-sighted new framework would prove to be more propitious.