Kenya now has one of its clearest legal statements on what happens when AI helps create a book, an image, or any other creative work. The Copyright Tribunal has ruled that AI-generated content cannot be copyrighted on its own.
A human has to have put in real creative effort for the work to qualify for protection. This was first spotted by Business Daily.
The case in question involves Aryeh Movement Limited and a woman named Cynthia Beldina Akoth. Akoth complained to the Kenya Copyright Board (KECOBO) in May 2025, saying Aryeh had registered certain literary works as its own without her permission. She said she was one of the actual creators.
KECOBO looked into it and sided with her, at least partly. The board told the two parties they needed to sit down and agree in writing on how to split the copyright ownership, and threatened to cancel the registration if that didn’t happen within seven days.
Aryeh pushed back and took the matter to the Copyright Tribunal, arguing that KECOBO had no business making that kind of call. KECOBO isn’t a court, and deciding who owns what and who counts as an author isn’t something the law lets the board do on its own.
The Tribunal sided with Aryeh on this point. It found that KECOBO acted outside its legal authority when it tried to resolve questions of authorship and ownership. Registering a work with KECOBO does not by itself decide who owns the copyright.
Under Kenyan law, copyright exists automatically the moment a work is created, and registration is just a formality that doesn’t need to happen at all for the rights to exist.
Because of this, the Tribunal set aside KECOBO’s decision entirely.
Can AI-Made Work Be Copyrighted?
While looking at the case, a separate question came up. Aryeh had described the disputed works as jointly authored, with Akoth listed as a script writer and as an “AI-image illustrator” who used AI-generated images.
That detail pushed the Tribunal into new territory: does Kenyan copyright law recognize works made with the help of AI?
The Copyright Act doesn’t mention AI at all. However, the Tribunal pointed to Section 22(3) of the Act, which says a literary or artistic work only qualifies for copyright if enough effort was put into giving it an original character, and if the work was actually written down or recorded in some fixed form.
Applying that rule to AI, the Tribunal said an AI-generated work is not automatically eligible for copyright.
READ: Anthropic Ordered to Pay $1.5 Billion in Largest AI Copyright Settlement
For it to qualify, the person claiming authorship has to show there was real human input and creative effort behind it, work that goes beyond just typing a prompt and using whatever the AI produces.
Importantly, the Tribunal did not rule on whether the specific books in this case actually meet that bar. Neither Aryeh nor Akoth submitted the actual works as evidence, and KECOBO never handed over its copies either.
Without seeing the works, the Tribunal couldn’t judge how much of them was human-made versus AI-generated. It said that kind of assessment needs to happen case by case, based on actual evidence.
What Other Countries Are Doing
The Tribunal looked at how other legal systems handle this question, and the pattern is fairly consistent.
In the United States, a federal court ruled in the Thaler v. Perlmutter case that an image created entirely by an AI system, with no human involved in the creative process, cannot be copyrighted because US law requires a human author.
The US Copyright Office has also said that when AI generates part of a work, the AI-made portions have to be disclaimed, and only the human-created parts can be protected.
The United Kingdom takes a slightly different approach. Its law has a specific category for “computer-generated” works, and it gives authorship to whoever made the arrangements necessary for the work to be created, even if there was no traditional human author involved.
That’s a more generous standard than the US one.
In the European Union, the standard comes from a 2009 case called Infopaq, which says a work is protected when it reflects the author’s own intellectual creation.
Since that phrase implies a human mind behind the work, most legal analysts in Europe read this as excluding fully AI-made content from protection.
Kenya’s position, based on this ruling, lands closer to the US and EU approach: no meaningful human creativity, no copyright.
A Precedent Set for AI and Copyright
This is one of the first times a Kenyan tribunal has directly addressed AI and copyright, even though it stopped short of applying the rule to the specific case in front of it.
For anyone in Kenya using AI tools to write, illustrate, or otherwise create content, just generating something with AI doesn’t automatically earn legal protection. There has to be a demonstrable, human creative contribution on top of it.
The ruling also cleared up a separate but related point. KECOBO cannot unilaterally decide who owns or authored a work when there’s a genuine dispute between parties. That question belongs to the Tribunal or the courts, not the registration board.
