By Faith Chepchumba
Artificial Intelligence is rapidly transforming creative industries, from music and literature to film and visual arts. In Kenya, as in many parts of the world, this technological advancement presents opportunities and challenges, particularly in intellectual property. AI systems capable of generating music, art, and written content blur the lines of authorship, raising pressing questions about copyright, data privacy, and the adequacy of existing laws.
Copyright Protectable Works
Copyright recognises the intellectual creation of an author from the moment the work is created. As stipulated under section 22 of the Copyright Act, 2001, copyright does not extend to the ideas themselves, but to the expression of those ideas.
To qualify for copyright protection, the protectable work has to meet two conditions.
- The work must be an original human intellectual creation
- The original human intellectual creation must be expressed in a manner identifiable with sufficient precision and objectivity.
The concept of protecting an expression as opposed to an idea is that an expression is more than a recording of work and is heavily linked to originality. Originality in the sense of the contribution of the author’s skill and labour. Thus, the Kenya Copyright Board will not register a copyright if it determines that a human being did not create the work. In cases where mixed human and non-human authorship is claimed, it is safe to say that copyright protection is available only for the human-authored selection and arrangement of the elements of the work.
Copyright and AI Infringement
The rise of AI in music creation is triggering major legal battles. In June 2024, the Recording Industry Association of America (RIAA) filed two lawsuits: one in Massachusetts against Suno (UMG Recordings, Inc. v. Suno, Inc., Case No. 1:24-cv-11611) and another in New York against Udio (UMG Recordings, Inc. v. Uncharted Labs, Inc., Case No. 1:24-cv-04777).
These platforms use AI to generate full songs in seconds based on user prompts. But there’s a catch: to function, the tools are trained on vast libraries of copyrighted music across genres, eras, and artists without consent or payment. Plaintiffs argue that this amounts to mass copyright infringement.
Imagine asking an AI: “Make me a song that sounds like Taylor Swift with a Selena Gomez-style verse.” Within seconds, it delivers a track mimicking their style without crediting or compensating the original artists.
The outcome of these cases will set crucial precedents for balancing innovation with protecting creators’ rights in the AI era.
According to the US Copyright Office report, using a machine as a tool does not negate copyright protection. Still, the resulting work is copyrightable only if it contains sufficient human-authored expressive elements, which can be read in full here. Additionally, the US copyright Office statement policy published in 2023, stated that“If a human submits a prompt to an AI system asking it to create a work, but the AI system has full control over the expressive elements of its output, such work is simply not the product of human authorship and is not copyrightable.”
Fair Use and AI: A Defence
The concept of “fair use” (or “fair dealing” in Kenya) is a defence that allows the limited use of someone else’s copyrighted material without permission under certain conditions. If AI-generated content falls within the scope of fair use. Companies may legally use copyrighted works to train AI models. If not, authorisation must be obtained, often requiring licensing fees.
In Kenya, the landmark Supreme Court case Communications Commission of Kenya & 5 others v Royal Media Services Limited & 5 others (CCK Case) established a six-factor test to assess fair dealing:
- The purpose of the dealing
- The character of the dealing
- The amount of the dealing
- Alternatives to the dealing
- The nature of the work
- The effect of the dealing on the work
While the Kenyan courts have not ruled on generative AI, the US Case Thomson Reuters Enter. Centre GmbH v. Ross Intelligence Inc., 694 F.Supp.3d 467. provides insight on how fair use might be interpreted.
The court ruled that ROSS Intelligence’s use of Thomson Reuters’ copyrighted legal material to train an AI system was not fair use. When examining the fair use factor, the purpose, and the character of the use, the courts found that the use did not have a further purpose or character different from Thompson Reuters, as ROSS was using Thompson Reuters headnotes as AI data to create legal research.
Regarding the nature of the work, the courts ruled in favour of ROSS, stating that the notes were not that creative and that the more creative the work was, the more protection it received.
On the amount of the dealing, the courts again favoured ROSS because ROSS did not make the copied material available to the public but instead used it for internal training. What matters is the amount and substance in making the copy, and if it has been made accessible to the public, which may serve as a competing substitute.
On the effect of the value, the courts favoured Thomson Reuters because ROSS’s product was designed to compete with it as a market substitute.
This case illustrates that even when some fair use factors are satisfied, significant market impact or lack of transformation can outweigh them.
Data Privacy and AI Training
Beyond copyright, data privacy raises its head. AI often learns from vast datasets, potentially containing personal information. How this data is collected, used, and protected requires scrutiny and robust regulations. While Kenya’s Data Protection Act, 2019 doesn’t directly regulate AI, it does require organisations to protect personal data by design. This means systems must be built to collect and use only the personal information needed for a specific task by default and nothing more.
Conclusion
The rise of artificial intelligence as a music creation tool opens a Pandora’s box of legal and ethical challenges, particularly regarding copyright protection in Kenya. Currently, works like those generated by AI exist in a legal no-man’s-land, lacking explicit protection. While labelling AI as an “author” might seem intuitive, it opens a can of worms regarding royalty distribution and legal liabilities.
With the rise of AI, the emerging technologies point to the unfitness of the current copyright Act to meet the challenges posed by AI. Thus, forcing a change in the legal system to define what constitutes authorship and how to protect human-generated content.
As AI continues to evolve, it will be interesting to see how the law and policy evolve to match it.
Leave a Reply