Who Owns AI-Generated Art? A Copyright Conundrum

The rise of Artificial Intelligence (AI) has sparked an artistic revolution, with AI systems now capable of generating breathtaking images, captivating music, and intricate literary works. But as these digital masterpieces proliferate, a fundamental question emerges: Who owns AI-generated art? This isn’t just an academic debate; it delves into the core principles of Intellectual Property (IP) law and challenges traditional notions of creativity and authorship.

The General Principles of Copyright Law: The Human Authorship Requirement

Traditionally, IP law, particularly copyright, is designed to protect the creations of human minds. Copyright grants creators exclusive rights to reproduce, distribute, perform, and display their original works. The underlying rationale is to incentivize human creativity and innovation by providing a framework for creators to benefit from their efforts. Key to this framework is the concept of “authorship,” which has historically been synonymous with human endeavor.

For a work to be eligible for copyright protection, it typically needs to meet certain criteria, including:

• Human Authorship: It needs human craftmanship.
• Originality: The work must be independently created and possess a modicum of creativity. It shouldn’t simply be a copy of another work.
• Fixation: The work must be expressed in a tangible medium.

However, these human-centric principles encounter a significant hurdle when confronted with AI-generated content.

The Complexities of AI-Generated Content Ownership

The difficulty in assigning ownership to AI-generated art stems from several factors:

• Lack of Human Authorship: If an AI system autonomously generates a work with minimal or no human intervention, can it be considered an “author”? Current IP laws generally do not recognize AI as having legal personality or the capacity to hold rights.
• The “Tool” vs. “Creator” Debate: Is AI merely a sophisticated tool, like a paintbrush or a camera, with the human user remaining the author? Or does its ability to generate content with little direct human control elevate it to a more autonomous “creator” status?
• Training Data and Infringement Risks: AI models are trained on vast datasets, often containing copyrighted material. This raises concerns about potential copyright infringement if the AI’s output is deemed derivative of the training data.
• Multiple Stakeholders: Who truly has a claim? The AI developer who coded the system? The person who curated and fed the training data? The user who provided the prompts? The owner of the AI system?

Recent Copyright Legislations and Judicial Precedence

Jurisdictions worldwide are grappling with these questions, with varying approaches and interpretations emerging. Here are some trajectories being taken by various jurisdictions:

The United States: Emphasizing Human Authorship

The U.S. Copyright Office has consistently maintained that human authorship is a prerequisite for copyright protection. Their guidance indicates that works generated entirely by AI, without human creative input, are not eligible for copyright.

A landmark case illustrating this stance is Thaler v. Perlmutter. Stephen Thaler sought to register a copyright for an image created by his AI system, the “Creativity Machine,” listing the AI as the sole author. Both the U.S. Copyright Office and subsequent court rulings, including the D.C. Circuit Court of Appeals, denied the application, affirming that the Copyright Act requires a human author. The court explicitly stated that “authors are at the center of the Copyright Act.”

However, the U.S. Copyright Office has clarified that if a human provides “significant creative input” such as editing, arranging, or selecting AI-generated elements, those human-contributed portions might be eligible for copyright. The key lies in the “level of control exerted by human creators.”

The European Union: “Author’s Own Intellectual Creation”

The European Union’s copyright framework also leans heavily on the concept of human authorship. Works must be the “author’s own intellectual creation,” reflecting their personality and resulting from their “free and creative choices.” This generally implies the necessity of a human author.

While the EU’s Artificial Intelligence Act primarily focuses on regulating AI systems based on risk, it includes transparency requirements for AI-generated content and mandates disclosures regarding the use of copyrighted data for training AI models. This doesn’t directly address ownership but facilitates identifying AI involvement and aims to address concerns around training data.

Kenya: A “Person by Whom the Arrangements Were Undertaken”

Kenya’s legal landscape offers a slightly different perspective, particularly under the Copyright Act. While Kenyan courts have not yet specifically ruled on the copyrightability of purely AI-generated work, the Act’s definition of “author” in relation to “a literary, dramatic, musical or artistic work or computer program which is computer generated” states it means “the person by whom the arrangements necessary for the creation of the work were undertaken.”

This phrasing opens the door for interpretation. It could attribute authorship to the user who makes the arrangements for the AI system to create the work, even if it involves minimal input like typing prompts. This suggests a potential recognition of the user’s role in initiating and guiding the AI’s creation.

The Copyright (Amendment) Act of 2022 primarily focused on revenue sharing for ring back tunes and establishing a National Rights Registry. While significant for the creative industry, it didn’t directly address the complexities of AI-generated content ownership. The Kenya Copyright Board (KECOBO) and other relevant bodies are actively monitoring international developments and engaging in discussions on the implications of AI for intellectual property law, and keeping up with trends will be vital.

The Way Forward: Navigating the New Frontier

The question of who owns AI-generated art is far from settled. The rapid advancements in AI technology constantly challenge existing legal frameworks, which were designed for a human-centric creative world.

Several approaches are being considered globally:

• Legislative Reforms: Many argue for new legislation specifically tailored to address AI-generated content, defining authorship and ownership in this evolving landscape.
• Hybrid Models: Some propose “hybrid authorship” models where both human and AI contributions are acknowledged, potentially leading to shared ownership or new licensing structures.
• Contractual Agreements: For AI tools that allow users to retain ownership of outputs (like some tiers of MidJourney or OpenAI’s ChatGPT terms), contractual agreements will play a crucial role in defining rights between the user and the AI developer.

For Authors and Content Creators

While the legal landscape is still forming, particularly in Kenya and the African continent in general, the clear message from recent developments, both internationally and locally, is that human input remains paramount for copyright protection.

For creators leveraging AI, understanding the nuances of how your involvement shapes the copyrightability of your work is no longer optional—it’s essential. Are you merely prompting, or are you significantly shaping, selecting, and refining the AI’s output? The distinction can mean the difference between owning your creation and having it fall into the public domain.

Don’t leave the future of your artistic endeavors to chance. Our team of experienced and well-versed intellectual property lawyers is at the forefront of this evolving field. We can help you navigate the complexities of AI and copyright, assess the copyright-ability of your AI-assisted works, and advise on best practices to safeguard your intellectual property in this new digital age.

Contact us today for a consultation. Let’s work together to protect your creativity and ensure your voice, and your art, are recognized and rewarded.