Generative AI and Copyright: Who owns creativity?

The question of whether works produced by generative AI deserve copyright protection is a complex and multifaceted one.

Artificial Intelligence has reached unforeseen levels of competence in recent times, with AI systems now capable of performing complex functions autonomously without the presence of a human agent or stimulus. Revenue from AI-Generated works is estimated to grow to $126 billion by the year 2025.

One of the most exciting and recent developments is the capability of AI to generate art using written prompts fed by users. As a result, a fundamental question arises pertaining to the authorship and ownership of the content so produced. Conflicting opinions exist in this regard, both in law and the scientific and legal communities.

The question of whether works produced by generative AI deserve copyright protection is a complex and multifaceted one.

Art And AI

The Merriam-Webster dictionary defines the term “Art” as “the conscious use of skill and creative imagination, especially in the production of aesthetic objects”. Where AI-Generated Art is concerned, content is generated by an expert system that “analyses countless works of art based on artistic style and produces a similar output.”

Such systems involve four key elements including “Inputs, Learning Algorithms, Trained Algorithms and Outputs”, with the final element being the tangible product commonly termed AI-Generated Art. One of the most prominent players in this field is the company OpenAI and their AI systems including DALL-E and more recently, Chat-GPT.

DALL-E and Chat-GPT use prompt or image-based user inputs to generate realistic content such as complex responses and artwork using an array of machine learning and neural network algorithms. As per the terms set by OpenAI for DALL-E, users “get full usage rights to commercialise the images you create with DALL·E, including the right to reprint, sell, and merchandise – regardless of whether an image was generated through a free or paid credit.”

However, OpenAI retains ownership and copyright of all generated content, or “Generations”, to the extent allowed by the law in force at the time.

What Copyright Laws Say

Legislations governing such works in countries such as India and the United States, though at a nascent stage, however, provide differing perspectives.

Copyright Law is primarily based on the utilitarian theory, which considers the purpose of copyright law to be an incentive for authors to “create expressive works for the public benefit.”  Copyright is granted to an “original work of authorship fixed in any tangible medium.”

The US Copyright Office has established a ‘Human Authorship Requirement,’ which explicitly states, “The US Copyright Office will register an original work of authorship, provided that the work was created by a human being.” It also provides that “the Office will refuse to register a claim if it determines that a human being did not create the work.”

Since AI-Generated Art by its very definition is created by the AI and not a human author, such works may be incapable of being registered as per the above guidelines. Accordingly, the US Copyright Office recently ruled that AI Art cannot be copyrighted and refused to register a work created by AI. As per the ruling, the image “lacks the human authorship necessary to support a copyright claim.”

This follows previous rulings such as that in Naruto v Slater, wherein it was ruled that a non-human applicant (in this case, a monkey) was not eligible for copyright protection and could not sue for copyright infringement. Earlier, in 1984, the Copyright Office released a ‘Compendium of Copyright Office Practices’ wherein it was stated that “for a work to be copyrightable, it must owe its origin to a human being.”

Laws Must Evolve

In fact, the latest practice compendium released by the Copyright Office, states that “the Office will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author.”

The above definition, however, presents an interesting perspective. Some may argue that since the generated works would not exist without the presence of a human agent who inputs the prompt, it would fall within the ambit of “creative input or intervention from a human author” and entitle the work to protection.

However, AI such as Chat-GPT operates by parsing and scraping information off of millions of web pages stored on the internet. There is a distinct possibility, however miniscule, that the output so produced may be substantially similar to an existing work, leading to a risk of plagiarism. Such AI has also been observed to generate wholly incorrect outputs, which poses a risk of spreading misinformation or fake news.

With the passage of time, AI such as Chat-GPT and  DALL-E will proliferate to the extent that numerous jobs and even entire professions may be rendered obsolete. It is the need of the hour to implement robust and comprehensive legislation around areas such as copyright and AI, to ensure that the power of AI is harnessed for good without otherwise leading to dire and undesirable consequences.

Rodney D Ryder is Founding and Senior Partner with Scriboard, a full service law firm with an intellectual property, technology and media law practice. He is the Co-Author of “Artificial Intelligence and Law: Challenges Demystified”. Views are personal and do not represent the stand of this publication.