Copyright Protection of AI-Generated Works in Kenya

A. Copyright protection of AI-generated works: What are the issues

Copyright law has traditionally extended automatic protection to original works that reflect human intellectual effort and are affixed in a material form.

Advances in technology, particularly artificial intelligence, are increasingly shaping Kenya’s creative economy. The use of AI in producing literary, artistic, and commercial content is expected to grow significantly, giving rise to three distinct scenarios, namely:

• Fully AI-generated works: Where the output is created entirely by the machine without human input.
• Hybrid works: Where AI-generated content contains distinguishable human-made elements.
• AI-assisted works: Where a human author creates the work with the aid of AI as a tool.

In the three scenarios, AI tools are creating or are used to create literary, artistic, sound, audio-visual, and other creative works. Under section 22 of Kenya’s Copyright Act, 2001, such works would ordinarily fall under works eligible for copyright and would traditionally qualify as protected works under the system of protection. The process of machine learning, data processing, and automated decision-making that produce these outputs, and the very nature of the outputs, challenge the copyright protection ecosystem in the following main ways:

• They put pressure on notions of human authorship, originality, ownership, and expression in material form.
• They put pressure on the economic and moral rationales for copyright protection, leading to legal grey areas on whether the copyright laws stimulate or impede technological
• While advances in technology promise to stimulate economic development, they also carry the risk of accelerating and expanding the transnational reach of copyright infringements, thereby creating significant challenges for enforcement.

From a regulatory standpoint, these impacts raise three important questions about the scope of copyright protection. Milczarek (2024) has distilled them into the following: Can AI creativity be treated as a ‘work’ within the meaning of copyright law? If the creativity generated by AI qualifies as a work of authorship, to whom do the rights to this work belong? Should the copyright protection system evolve to extend protection to AI creations as well?

B. Copyright protection of AI-generated works: What are the issues in Kenya?

Kenya’s Copyright Act, 2001 recognizes copyright in computer programs as a means of expressing protected works, a mechanism for copying works, and a form of literary work eligible for copyright protection.
However, the Kenya‘s Copyright Act, 2001 does not expressly address the copyrightability of AI-generated, machine-generated, or computer-generated works. Consequently, questions relating to the ownership and protection of AI-assisted works continue to be determined by reference to the existing originality requirements under the Copyright Act, 2001 as well as emerging jurisprudence.

C. Jurisprudence in Aryeh Movement Limited v Akoth & Another

1. Scope of the dispute
   Aryeh Movement Limited v Akoth & Another was an appeal case filed at the Kenyan Copyright Tribunal (the Tribunal). The appeal challenged a decision of the Kenya Copyright Board (KECOBO) concerning registration of certain literary works. The principal dispute concerned authorship and ownership of the registered works, and the Copyright Tribunal observed that parts of the works involved AI-generated images that had been curated and adapted by one of the parties. This prompted the Tribunal to consider a subsidiary issue: whether AI-generated content is eligible for copyright protection under Kenyan law, and the extent to which human involvement affects copyright eligibility.
2. What the Tribunal decided
   The Tribunal observed that the Copyright Act does not specifically provide for AI-generated works. Accordingly, the applicable legal standard remains the originality test under Section 22(3) of the Act. This section requires that sufficient effort must have been expended in creating a work to give it an original character and the work must have been reduced into material form. The Tribunal declined to determine whether the disputed literary works were themselves copyrightable, noting that the parties had neither produced the actual works nor adduced sufficient evidence to establish the extent of AI involvement or human creative contribution.

Nevertheless, the Tribunal examined comparative judicial approaches to the copyrightability of AI-generated works, considering developments in three jurisdictions:

• The United States’ human-au­thorship requirement, as applied in Thaler v Perlmutter;
• The United Kingdom’s recogni­tion of computer-generated works under Sections 9(3) and 178 of the Copyright, Designs and Patents Act 1988; and
• The European Union’s ‘author’s own intellectual creation’ stan­dard articulated in Infopaq Inter­national A/S v Danske Dagblades Forening.

Consequently, the Tribunal clarified that AI-assisted works are not auto­matically excluded from copyright protection. The Tribunal further ruled that the eligibility will depend on the degree of human authorship and creative control exercised in producing the final work, as required under Section 22(3)(a) of the Kenya‘s Copyright Act, 2001.

3. Implications for copyright protection in Kenya in the era of AI

The Tribunal’s decision does not establish a blanket rule excluding AI-generated or AI-assisted works from copyright protection. Instead, it clarifies that eligibility will depend on whether sufficient human inter­vention or creative effort exists to satisfy the originality requirements under the Copyright Act. Purely machine generated works, without meaningful human input, are gener­ally not protected. This reinforces the anthropocentric approach in Kenyan law, where human author­ship remains central to determining copyrightability. While the Tribunal’s inability to de­termine the copyrightability of the disputed literary works was argu­ably a missed opportunity, the case nonetheless underscores the impor­tance of human authorship. Copy­right protection requires demon­strable human effort that imparts originality to the work.

The Tribunal’s reasoning also offers practical guidance on evidentiary considerations in future disputes involving AI-assisted works. When AI is deployed as a creative tool, au­thors and claimants must be pre­pared to demonstrate the extent of human creative contribution to the final output.

The reliance on comparative juris­prudence highlights that human involvement in directing, selecting, arranging, editing, or otherwise shaping AI-generated content will be a decisive factor in assessing copyrightability.

4. Law reform and future of protection of AI-generated works

From a legal reform perspective, the Kenyan Draft Copyright and Related Rights Bill, 2026 proposes signifi­cant changes to copyright adminis­tration and enforcement but does not expressly address the copy­rightability of AI-generated works.

Consequently, questions relating to AI-assisted content will likely contin­ue to be resolved by reference to the originality requirements under the Copyright Act, 2001 and the principles articulated by the Tribunal in
Aryeh case.

In parallel, the Artificial Intelligence Bill, 2026 introduces obligations relat­ing to transparency, accountability, and the labelling of AI-generated con­tent. While these provisions do not directly determine copyright owner­ship, they are expected to enhance the traceability of AI use in content creation. This may assist parties in demonstrating the extent of human involvement in AI-assisted works or in establishing AI’s role in copyright infringement, thereby providing greater certainty in future disputes. Until Kenya enacts legislation specifically addressing AI-generated works, the interpretation of the originality test under Section 22(3) of the Copy­right Act, 2001 in the Aryeh case will continue to offer a useful guide.

5. Conclusion and Recommendation

The decision of the Copyright Tribunal in the Aryeh case is Kenya’s first substantive judicial engagement with the question of copyright protec­tion in the age of AI. Its grounding in comparative jurisprudence and em­phasis on the centrality of human authorship under Section 22(3) of the Copyright Act, 2001, lay an important foundation for the evolution of Kenyan copyright law. As AI continues to reshape the creative economy, this decision signals that the law will adapt by focusing on demonstrable human input, ensuring that technological innovation is balanced with the enduring copyright protection principle of human originality.

Going forward;

• Publishers, media houses, advertising agencies, technology companies, and professional service firms that deploy AI should adopt internal policies gov­erning its use and put in place mechanism that can enable it to track and-demonstrate the extent of human creative input in AI-assisted works.
• Publishers, media houses, advertising agencies, technology companies, and professional service firms that deploy AI should maintain contractual safe­guards. Such measures will strengthen future claims to copyright protection and reduce uncertainty in case ownership disputes arise.
• Creators and businesses that routinely deploy AI should maintain clear records of the human input involved in producing AI-assisted works. Such documentation will be critical in establishing copyright ownership should disputes arise.