CAN AI BE GRANTED COPYRIGHT?

In recent developments, Asian News International (ANI) has filed a lawsuit against OpenAI, alleging copyright infringement for using ANI’s articles to train its AI models. This case has reignited the ongoing debate surrounding the complex relationship between artificial intelligence (AI) and copyright law. ANI claims that OpenAI unlawfully utilized its articles to train their large language models (LLMs) without permission, thereby infringing upon ANI’s copyright.

This is not the first instance in which AI and copyright have collided. As technology advances and AI systems gain increasing autonomy, the need to address intellectual property issues related to AI becomes more pressing. The issue of copyright and AI remains an ongoing legal saga with no clear resolution as jurisdictions around the world struggle to find a consistent approach to this complex matter.

Copyright in AI-Generated Works in Different Jurisdictions

One of the key questions is whether AI can be considered an author, as an author is typically regarded as the first owner of any copyrightable material. Another crucial question is whether AI-generated works are acknowledged for copyright. These unresolved issues have led to varying approaches in different countries.

In USA:

Under the Copyright Act of 1976, copyright protection is granted to "original works of authorship fixed in any tangible medium of expression." This means that for a work to qualify for copyright, it must be fixed in a tangible medium either by the author or under their authority. However, Act does not provide a specific definition of "author,". In case of Stephen Thaler v. Shira Perlmutter [Civil Ac on No. 22-1564 (BAH)], the issue was whether work autonomously generated by an AI system is copyrightable. ?The court relied on dictionary definitions of author, which describe an author as “one that is the source of some form of intellectual or creative work” and “the creator of an artistic work; a painter, photographer, filmmaker, etc”. Hence, the court interpreted that the originator of work must be a human in order to be considered as author and work generated by AI is not copyrightable.

In India:

Section 2 (d) (vi) of Indian Copyright Act, 1957 defines author for computer- generated work as, “in relation to any literary, dramatic, musical or artistic work that is computer-generated, the person who causes the work to be created”. In case of Tech Plus Media Private Ltd. vs Jyoti Janda [2014 (60) PTC 121 (Del)] have held that a author has to be “a natural person”. Hence, even in India, it is clear that a human individual, responsible for the creation of the work—such as the programmer or the person who directed the AI—would be regarded as the author and not the AI itself. In India, there was an attempt to recognize an AI system named "Raghav" (Robust Artificially Intelligent Graphics and Art Visualizer) as a co-author of an artistic work. However, this recognition was later withdrawn when the Copyright Office realized it had mistakenly granted the registration. This reinforces the position that AI cannot be granted neither sole nor joint authorship.

In UK:

The United Kingdom has acknowledged the concept of AI-generated work under its copyright law. Section 178 of the Copyright, Designs, and Patents Act (CDPA) defines a "computer-generated work" as one "generated by a computer in circumstances such that there is no human author." However, this definition must be read in conjunction with Section 9(3) of the CDPA, which designates the author of a computer-generated work as the person who undertakes the necessary arrangements for the creation of the work. Therefore, while the work itself may be generated by AI, the human responsible for directing or enabling the creation of that work is considered the author under UK law.

While USA does not acknowledge copyrightability of AI-generated work, the situation in India remains a grey area due to the lack of clear legal developments. In contrast, UK have taken a clear stance in acknowledging AI-generated works. Not granting acknowledgment to AI-generated work could hinder technological progress, and therefore, it must be acknowledged. While it has been established that AI cannot be considered an author, either solely or jointly with another person, the key question remains: should the ownership lie with the programmer, or with the individual who provides commands or directions to the AI for generation of such output?

The answer to this question is likely to depend on the specific facts of each case. Denial Gervais, a prominent scholar, has proposed a test to assess causation in AI-generated works. According to Gervais, the focus should be on determining whether the human's instruction or input directly influenced the AI's output, with courts evaluating the originality of the work.

A relevant case in this context is Nova Productions Ltd. v. Mazooma Games Ltd. [2006 RPC 379], where Nova claimed copyright in the frames generated by users in a video game. “The court ruled that the copyright should belong to the person who designed the game's appearance, devised the rules, and wrote the program, rather than the players who did not contribute any artistic skill or labor”.

The Current Uncertainty of AI and Copyright

In the current legal landscape, it is evident that artificial intelligence (AI) is not considered as author nor can it hold ownership over any work it generates. Consequently, ownership must reside with a person who either programmed the AI or the user who gave directions to produce the output. The determination of ownership will largely depend on the individual who exerted the most influence over the final output. In light of recent legal precedents and principles, it is clear that the question of who owns the copyright to AI-generated works remains a complex and evolving issue. Courts will need to continue to navigate this uncertainty, with decisions varying on a case-by-case basis.

Ownership of AI-generated works can be contractually determined between the programmer and the user. Copyright typically belongs to the creator of the work, but since AI cannot hold copyright, so human involvement is required. Contracts can clarify whether the programmer or the user will own copyright over the output, especially in paid versus unpaid versions of AI tools. Paid versions may offer broader licenses or full rights to the user, while free versions may retain more restrictive rights for the programmer. As AI and copyright law evolve, clear agreements are crucial for defining ownership and usage rights.

Furthermore, generative AI systems are trained on vast databases containing works from numerous artists, often publicly available, to draw inspiration, similar to the way a human artist might. Therefore, the use of these databases by AI can be considered fair use, as it mirrors the process of human creativity. However, if the output generated by the AI closely resembles an individual artist's work, that artist should be entitled to adequate compensation for any potential copyright infringement. This ensures that while AI can be used as a tool for creativity, the rights of original creators are still protected.

In conclusion, as AI technology advances, legal systems will need to evolve in response to the challenges presented by AI-generated works and copyright. While the question of AI's rights may remain uncertain for now, ongoing judicial intervention will play a crucial role in shaping the future of AI and intellectual property law.

?