Introduction
Artificial Intelligence (AI) is rapidly transforming Kenya’s creative economy, from publishing and advertising to music production and visual arts. Yet as AI tools become more accessible, a critical legal question has emerged: who owns the output, and can it be protected under copyright law? In 2026, Kenya’s Copyright Tribunal delivered one of Africa’s clearest judicial statements on this issue, ruling that AI-generated works cannot enjoy copyright protection unless a human author demonstrates sufficient creative effort and originality. This article examines how AI interacts with Kenya’s intellectual property framework, analyses the Tribunal’s recent jurisprudence, and explores the practical implications for creators, businesses, and policymakers navigating the intersection of technology and law.
Kenya’s Intellectual Property Framework
Kenya’s intellectual property (IP) regime is anchored in several statutes, with copyright governed primarily by the Copyright Act (Cap 130), administered by the Kenya Copyright Board (KECOBO). The Act protects seven categories of works: literary, musical, artistic, dramatic, audio-visual works, sound recordings, and broadcasts. Crucially, copyright protection in Kenya is automatic upon fixation of a work in material form—registration with KECOBO is voluntary and serves only as prima facie evidence of ownership, not a prerequisite for subsistence.
For a work to qualify for copyright, it must satisfy two core statutory requirements under Section 22(3) of the Copyright Act:
1. Originality: “Sufficient effort has been expended on making the work to give it an original character.”
2. Fixation: The work must be “written down, recorded or otherwise reduced to material form.”
These thresholds presuppose human agency. As the Copyright Tribunal recently emphasised, the phrase “sufficient effort” implies human intellectual labour, skill, and judgment—not mere mechanical or algorithmic output. This interpretive stance aligns Kenya with jurisdictions like the United States and the European Union, which similarly require human authorship for copyright eligibility, rather than the United Kingdom’s more permissive approach that recognises “computer-generated works” with authorship attributed to the person who made the necessary arrangements.
The Landmark 2026 Copyright Tribunal Ruling
The pivotal case is Aryeh Movement Limited v. Cynthia Beldina Akoth Okello (COPTA/E001/2025), decided in rulings dated July 9 and August 24, 2026. The dispute arose when Akoth complained to KECOBO that Aryeh had registered literary works—Bible scripture stories—as its own without her consent, despite her role as scriptwriter and “AI-image illustrator.” KECOBO intervened, directing the parties to negotiate a written agreement on copyright ownership or face expungement of the registration.
Aryeh challenged KECOBO’s directive before the Copyright Tribunal, arguing the Board had exceeded its statutory mandate by attempting to determine authorship and ownership—matters reserved for the Tribunal or courts. The Tribunal agreed, setting aside KECOBO’s decision as ultra vires (beyond legal authority).
Key Holdings on AI and Copyright
While the Tribunal declined to rule definitively on whether the specific works in dispute were copyrightable (due to lack of evidence), it made several groundbreaking obiter dicta (judicial observations) on AI-generated content:
- AI cannot be an author: Under Kenyan law, only natural or legal persons can hold copyright; AI lacks legal personality and therefore cannot be recognised as an author or first owner.
- Human intervention is essential: Works generated solely by AI—without meaningful human creative input—do not satisfy the “sufficient effort” test under Section 22(3)(a) and are ineligible for copyright.
- AI-assisted works may qualify: If a human exercises sufficient skill, judgment, or creativity in shaping, selecting, arranging, or refining AI output, the resulting work may attract copyright protection—but only to the extent of the human contribution.
- Burden of proof lies with the claimant: Anyone asserting copyright over AI-assisted content must demonstrate tangible evidence of their creative intervention beyond merely typing a prompt.
The Tribunal noted that neither party submitted the actual works as evidence, nor did KECOBO produce its copies, making factual assessment impossible. Nevertheless, the ruling establishes a clear human-authorship floor: no human creativity, no copyright.
Statutory Interpretation: Gaps and Ambiguities
Kenya’s Copyright Act does not expressly mention AI-generated works. However, Section 2 defines “author” in relation to computer-generated works as “the person by whom the arrangements necessary for [the work’s] creation were undertaken.” This provision, inherited from Commonwealth models, was designed for traditional software (e.g., code-written programs), not generative AI systems that autonomously produce content.
The Tribunal’s interpretation effectively narrows this definition: “arrangements” must involve creative control, not just technical setup. Mere procurement or deployment of an AI tool does not suffice; the human must contribute original expression.
This creates legal uncertainty for hybrid works. For instance:
- A novelist using AI to draft chapters but heavily editing and restructuring them may claim copyright in the final manuscript.
- A marketer generating social media images via prompt alone likely cannot.
The absence of a bright-line test means each case turns on factual inquiry—a costly and unpredictable process for creators.
Comparative Perspectives: Global Trends
Kenya’s approach mirrors emerging global consensus:
- United States: The Copyright Office and courts (e.g., Thaler v. Perlmutter) consistently deny registration to purely AI-generated works, requiring human authorship. AI-assisted works must disclaim AI-generated portions.
- European Union: The Court of Justice’s Infopaq standard requires works to reflect the “author’s own intellectual creation,” implicitly excluding non-human outputs.
- United Kingdom: Uniquely, the UK Copyright, Designs and Patents Act 1988 (Section 9(3)) grants authorship of computer-generated works to “the person by whom the arrangements necessary for the creation of the work are undertaken.” However, even here, debates continue over whether this applies to generative AI.
- India: Recently rejected an application to register an AI system (DABUS) as author, affirming that only humans can hold authorship, though the work itself may be original.
Kenya thus aligns with the restrictive, human-centric model, prioritising doctrinal coherence over technological accommodation.
Practical Implications for Creators and Businesses
The Tribunal’s ruling has immediate consequences for Kenya’s digital economy:
1. Content Creation and Publishing
Writers, illustrators, and designers using AI tools must document their creative process. Contracts should specify ownership of AI-assisted outputs, and registration applications must disclaim AI-generated portions, claiming only human-authored elements.
2. Marketing and Advertising
Agencies generating copy, images, or videos via AI cannot assume automatic copyright. To secure enforceable rights, they must ensure staff add substantial creative input—editing, curating, or combining AI outputs in original ways.
3. Software and Tech Development
While computer programs themselves are protected, AI-generated code or assets within software may not be. Developers should audit AI dependencies and maintain records of human oversight.
4. Enforcement Challenges
Without clear ownership, infringement claims become difficult. If a work lacks human authorship, it may be treated as public domain, allowing unrestricted use by competitors.
5. Investment and Licensing Risks
Startups relying on AI-generated content face valuation risks. Investors may demand warranties that outputs are human-authored or legally protectable.
Legislative and Policy Responses
Recognising these gaps, Kenya is undertaking IP law reforms:
- Copyright and Related Rights Bill 2026: Proposes treating “the person who made the arrangements necessary for the creation of the work” as the author of AI-generated works, potentially expanding protection. However, critics argue this could undermine the human-creativity principle unless “arrangements” are strictly defined.
- National Artificial Intelligence Strategy 2025–2030: Acknowledges AI’s transformative potential but lacks specific copyright guidance, calling for future legislative amendments.
- Presidential Directive (August 2025): Orders KECOBO, the Kenya Law Reform Commission, and the Attorney General to overhaul the Copyright Act to address emerging technologies.
These initiatives signal awareness but stop short of resolving core ambiguities. A balanced approach would:
- Clarify the threshold for “sufficient human intervention.”
- Introduce a sui generis (unique) right for non-original AI outputs (e.g., database-like protection).
- Mandate disclosure of AI use in registration applications.
Enforcement Mechanisms and Remedies
Despite uncertainties, Kenya’s enforcement framework remains robust:
- KECOBO: Can conduct raids, seize infringing copies, and mediate disputes (though not determine ownership).
- Copyright Tribunal: Hears appeals on registration, licensing, and CMO (Collective Management Organization) tariffs.
- High Court: Adjudicates infringement claims, granting injunctions, damages, or accounts of profits.
- Criminal Sanctions: Willful infringement attracts fines up to 10 times the work’s market value or imprisonment up to 10 years (20 years for repeat offenders).
However, enforcement presupposes valid copyright. If a work is deemed AI-generated without human input, no cause of action exists.
Case Study: The Aryeh–Akoth Dispute
The facts illustrate real-world complexities:
- Akoth claimed authorship of scripture stories, acknowledging use of AI for illustrations.
- Aryeh registered the works as commissioned, asserting ownership.
- KECOBO ordered a written agreement, overstepping its mandate.
- Tribunal voided KECOBO’s directive but affirmed that AI portions likely lacked protection.
Outcome: Both parties potentially lost enforceable rights due to inadequate evidence of human creativity. Lesson: Document your contribution.
Recommendations for Stakeholders
For Creators:
- Maintain drafts, edit logs, and version histories showing human input.
- Use contracts to allocate ownership of AI-assisted works.
- When registering, clearly delineate human vs. AI contributions.
For Businesses:
- Audit AI workflows for copyright compliance.
- Train staff on “creative threshold” requirements.
- Include IP warranties in vendor and freelancer
For Policymakers:
- Amend the Copyright Act to define “sufficient human intervention.”
- Consider a limited neighbouring right for AI outputs (e.g., 5-year term).
- Issue KECOBO guidelines on AI disclosure in registrations.recordoflaw+1
For the Judiciary:
- Develop precedent on factual tests for originality in AI contexts.
- Encourage expert testimony on AI capabilities and limitations.
Conclusion
Kenya’s Copyright Tribunal has drawn a firm line: copyright protects human creativity, not machine output. While this preserves doctrinal integrity, it leaves a grey zone for the growing class of AI-assisted works. As the National AI Strategy unfolds and legislative reforms advance, Kenya has an opportunity to craft a nuanced framework that incentivises innovation without diluting authorship. Until then, creators must tread carefully—ensuring their AI collaborations remain firmly anchored in human ingenuity.
The message is clear: in Kenya’s eyes, AI is a tool, not an author
Shafiq Taibjee
Lawyer/Arbitrator/Mediator/Certified Islamic Arbitrator/Court Annexed Mediator/Banking Mediator. Honorary Fellow IICRA (UAE)