A walk through the Copyright Act, what it means for the solo developer working alone, and what the recent High Court judgment in Peter Nthei Muoki and Beluga Limited v Safaricom PLC can teach the public about how copyright actually works.
Why this matters
Every law that protects creators is also a law that limits what others can build on top of. That is the bargain at the heart of intellectual property.
Make the protection too narrow and creators give up before they begin. Make it too broad, and every person who later works in the same field is exposed to claims they cannot foresee or defend against.
Either error closes a door that should have stayed open. The whole system depends on the line being drawn carefully, in the right place, for the right reasons.
When a court applies that line in a way the statute does not authorise, the consequences travel far beyond the parties in the courtroom.
Every future innovator, every business that engages with outside ideas, every Kenyan trying to understand what they own and what they do not — all of them are affected by how that line gets drawn.
Copyright is not abstract. It is the framework that decides whether the next person who has a good idea will be protected, or sued, or simply ignored.
The recent judgment in Peter Nthei Muoki and Beluga Limited v Safaricom PLC provides a useful occasion to walk through the Kenyan Copyright Act and let it speak for itself.
What the court awarded
In a judgment delivered virtually on 8 May 2026, Justice Josephine Mong'are made the following orders.
She declared that Safaricom had violated the plaintiff's rights under the Copyright Act. She awarded general damages of Sh 1,400,067,000, calculated as one percent of M-Pesa's total revenue for financial year 2024, with interest at court rates from the date of judgment until payment in full.
She further ordered Safaricom to pay 0.5 percent of M-Pesa's gross revenue every financial year, commencing 31 March 2025, for as long as Safaricom continues to operate the Manage Child Account or M-Pesa Go product or any substantially similar parent-child control functionality.
She framed this ongoing payment as a compulsory licence in lieu of a permanent injunction.
She declined to grant an injunction, reasoning that one would disrupt millions of users.
What copyright protects in Kenya
The Copyright Act, Chapter 130 of the Laws of Kenya, is the governing statute. It protects literary works, musical works, artistic works, audio-visual works, sound recordings, and broadcasts. Within the definition of literary works, the Act expressly includes computer programs and compilations.
So code is protected, but it is protected as a literary work — in the same way a novel or an essay is protected. The protected thing is the set of instructions you wrote, not the task the program performs.
Two things are required for protection to exist.
Originality. Sufficient effort must have been put into the work to give it an original character. The threshold is not high — the law does not ask whether the work is novel, inventive, or unprecedented, only whether genuine effort has gone into it.
Fixation. The work must be reduced to a form that can be perceived and reproduced — written down, recorded, saved, committed. Both requirements are quietly satisfied the moment you save your work somewhere.
Ownership of the copyright belongs to the author — the person who actually created the work — unless the work was created in employment, in which case it belongs to the employer, or unless it was assigned to someone else by written agreement.
The owner has the exclusive right to control reproduction, translation, adaptation, distribution, public performance, broadcasting, and communication to the public of the work. These are the rights that, when exercised by someone else without permission, amount to infringement.
Beyond the Kenyan Act, Kenya is bound by the international framework. The TRIPS Agreement, which Kenya joined through the World Trade Organization in 1995, sets a rule every member country must respect:
Copyright protection extends to expressions, not to ideas, procedures, methods of operation, or mathematical concepts.
Computer programs are protected as literary works under the Berne Convention — meaning the protection is for the code, not for what the code does.
This is the idea-expression dichotomy. It is the foundation of copyright everywhere, and it is the line on which everything else depends.
Proving originality and fixation
Two requirements decide whether your work qualifies for copyright protection at all. Both can be quietly satisfied in the course of normal work, but knowing what they look like in practice helps when a dispute arises.
Originality is shown by evidence that the work reflects your own intellectual effort rather than mechanical copying.
The threshold is not high — the law does not ask whether the work is novel, inventive, or unprecedented, only whether sufficient effort has been expended on it to give it an original character.
In practice, originality is proved through artefacts of the creative process: drafts, iterations, design notes, commit histories showing the work taking shape over time, comments and decisions that reveal the thinking behind the code, evidence of the problems you considered and the choices you made.
A repository with a long history of incremental commits, each reflecting work in progress, is much stronger evidence of originality than a single dump of finished code with no trail behind it.
Fixation is shown by evidence that the work exists in a form that can be perceived and reproduced.
Saving the file is fixation. Committing to a repository is fixation. Emailing the work to yourself is fixation.
The law does not specify a particular medium; it requires only that the work has a tangible, durable form.
What proves fixation is the artefact itself — the file, the commit, the timestamped record — together with metadata showing when it came into existence and in whose hands.
Two habits strengthen both proofs.
Commit your work to version control regularly, with meaningful messages.
Where the work has real commercial value, register it with the Kenya Copyright Board so that you have a dated certificate confirming the form the work was in on a specific date.
The commit history shows originality through the visible labour of creation. The registration shows fixation by anchoring the work to a specific point in time under your name.
Neither step creates the copyright. The copyright is yours from the moment you write the work. Both steps make the copyright provable.
Proving infringement
To win a copyright case, a claimant must show three things.
Subsistence. The work qualifies for copyright — it is original and fixed.
Ownership. The claimant is the author of the work, or has had the rights assigned to them.
An infringing act. The defendant has done something only the owner is allowed to do — copied the work, adapted it, distributed it, or otherwise exploited it without permission.
The standard test for proving the infringing act has two parts.
Access. The defendant had the opportunity to see the work.
Substantial similarity. What the defendant produced reproduces a substantial part of the protected work.
Substantial similarity is measured against the protected expression, not against the underlying idea. This is the line on which most software cases turn.
In software cases the rule has been settled clearly in Navitaire v easyJet and SAS Institute v World Programming: copying the functionality of a program — without copying the source code or its distinctive structural elements — is not copyright infringement.
You can build a program that does the same thing, in the same ways, as someone else's program. What you cannot do is copy the code itself.
If you prove infringement: what you gain
If you prove your case, the law gives you a structured set of remedies.
Damages. Money to compensate for the infringement.
The amount is calibrated by what you have lost — the income or value you would have received had the infringement not happened — or by what the defendant has gained from the infringement specifically.
Where the infringement is flagrant and ordinary measures would not provide effective relief, additional damages are available.
The common thread is that damages are calibrated to the wrong itself and to the specific protected work — not to the defendant's overall commercial activity in unrelated areas.
An injunction. A court order requiring the defendant to stop using the protected work. This is the strongest remedy because it ends the infringement going forward.
Damages in lieu, or a defined-term licence. If the court refuses an injunction on equitable grounds — for instance, because stopping the use would harm third parties — the orthodox alternatives are a one-off sum reflecting the value of future use, or a licence on agreed terms for a defined period.
Costs. The successful party recovers their costs of the litigation.
What is not generally available is a court-imposed compulsory licence — a perpetual obligation requiring the defendant to keep paying for ongoing use without your consent.
Compulsory licences in Kenyan copyright law exist only in narrow, specific contexts written into the statute: translation rights for developing countries under the Berne Appendix, accessibility for persons with visual impairments under the Marrakesh Treaty, and certain disputes referrable to the Competent Authority.
Compulsory licensing is a creature of statute, not a general judicial power.
How the judge reasoned through Muoki v Safaricom
Before assessing the judgment, it helps to follow how the trial court arrived at its conclusions.
Step one: the plaintiff's documentary foundation.
The court accepted the KECOBO Certificate of Registration as prima facie evidence of ownership, the detailed written documentation as fixation, and the Source Code Repository History — which traced the work back to October 2019 — as corroboration of timeline.
Subsistence, ownership, and a chronology predating Safaricom's claimed independent development were all established.
Step two: access.
The plaintiff approached Sylvia Mulinge in March 2021 and shared detailed insights with Sitoyo Lopokoiyot between March and June 2021. They met on 22 June 2021.
Safaricom did not call Mr Lopokoiyot to deny the meeting. Access was treated as uncontroverted.
Step three: testing Safaricom's competing origin stories.
Safaricom offered two accounts of how its product came to be. The court rejected both.
The first — a verbal request from the CBK Governor about minors and betting — was dismissed on the basis that a telecom would not act on undocumented verbal instructions, and that advising on product design is not CBK's statutory role.
The second — that Huawei had independently developed the product from September 2020 — collapsed under the evidence.
Safaricom never produced the final Functional Requirements Specification. Its witness could not say when Huawei produced the final product. Huawei's witness conceded there was no formal instruction or directive from Safaricom to commence development.
The judge concluded the Huawei proposal was a "belated attempt to create a paper trail to defeat the Plaintiffs' claim."
Step four: inferring copying.
With both origin stories rejected, the court applied the Designers Guild method: where similarities justify an inference of copying and the defendant's evidence of independent provenance is rejected, the finding of copying carries the substantiality requirement.
The temporal sequence reinforced this. Disclosure was in March-June 2021. The product was test-run shortly after and launched in November 2022.
Step five: identifying what was taken.
This is the substantive heart of the judgment.
The court distinguished the plaintiff's claim from a monopoly over an idea. The concept of parental control, the court accepted, is unprotectable.
What the plaintiff claimed exclusivity over, the court held, was a specific expression of that concept: the detailed USSD menu tree, the sequence of operations, the types of restrictions, the reporting mechanisms, and the system responses in PExhibit 6.
Safaricom's product, the court found, reproduced those details. That, in the court's view, is what copyright protects.
Step six: the remedies.
The court refused a permanent injunction on public interest grounds, reasoning that millions now rely on the functionality.
On damages, it rejected Safaricom's submission that only nominal damages should follow, noting that Safaricom "did not seek a licence, they simply took it and the Plaintiffs were deprived of a negotiating opportunity."
It examined Safaricom's reported M-PESA revenues — KSh 82.6 billion in FY2021 (the year of disclosure), rising to KSh 140 billion in FY2024 — and awarded damages of 1% of FY2024 M-PESA revenue.
In place of the refused injunction, the court ordered a perpetual 0.5% royalty, framed as a compulsory licence.
The judge closed by observing that the case is "a cautionary tale for innovators and corporations alike."
That is the reasoning. The questions it raises follow.
How does this apply in Muoki v Safaricom?
The framework above is what the law of copyright in Kenya looks like in its ordinary operation. Reading the recent judgment with that framework in mind raises a question worth asking carefully: was this, in the strict sense, a copyright case?
On the finding of infringement.
The court did not stop at the language of "uncanny resemblance and functionality." It identified specific elements in the plaintiff's documentation — the USSD menu tree, the sequence of operations, the types of restrictions, and the reporting mechanisms — and found that Safaricom's product reproduces those details.
That is the doctrinally important step, and it is on the record.
The harder question — the one an appellate court would have to confront — is whether those elements are protected expression, or unprotected functional structure described in expressive language.
The judgment relies on authorities that address how copying is proved (Designers Guild v Russell Williams, the local Franz Frederichs) and on authorities affirming the idea-expression dichotomy in general terms (Solut Technology v Safaricom, Jack J. Khanjira v Safaricom). Each of these is good law.
What the judgment does not cite, and what an appeal would need to engage with, is the line of jurisprudence that has developed specifically around the question of what constitutes protected expression in software cases.
Navitaire v easyJet and SAS Institute v World Programming settle the rule that functionality, programming logic, and the way a program performs its operations are not protected by copyright even when they can be described in detailed prose.
What is protected is the source code itself and specific expressive choices not compelled by the function.
An appellate court applying that line would have to ask a sharper question.
Would another team building a parent-child mobile money product, working under the same CBK regulatory requirements and the same USSD medium constraints, independently arrive at substantially the same menu tree, the same sequence of operations, the same types of restrictions, and the same reporting structures?
If yes, the similarity is largely the product of shared function, and the orthodox line places it outside copyright.
If no, the finding rests on protected expression and the orthodox line is satisfied.
This is the question on which a careful appeal would turn.
On the damages.
The court awarded Sh 1.4 billion, calculated as 1% of M-Pesa's total revenue of Sh 140 billion for financial year 2024.
The reasoning offered was that this sum was "a negligible cost to Safaricom" given the scale of its business and its financial capacity.
But damages in copyright are calibrated to the wrong itself — to what the claimant has lost, or to what the defendant has gained from the infringement specifically.
M-Pesa's total revenue includes peer-to-peer transfers, paybills, withdrawals, Fuliza, M-Shwari, international remittance, and many other functions unrelated to the parent-child account feature in dispute.
The orthodox method would apportion damages to the specific impact of the alleged infringement on the specific protected work. The methodology applied — anchored to the defendant's overall capacity to pay — is a different exercise.
On the perpetual royalty.
The court ordered Safaricom to pay 0.5% of M-Pesa's gross revenue every financial year, in perpetuity, framed as a compulsory licence in lieu of a permanent injunction.
But compulsory licences in Kenyan copyright law exist only in the narrow statutory contexts identified above — none of which applies here.
Where an injunction is refused on equitable grounds, the orthodox alternatives are damages in lieu (a one-off sum) or a defined-term licence on agreed terms.
A perpetual royalty on a defendant's gross revenue, ordered without a statutory anchor, is not among the remedies the Copyright Act provides.
A critical case to watch
If Safaricom takes this judgment to the Court of Appeal, the outcome will matter beyond the parties.
Three features of the trial court's decision would, if upheld, become part of Kenyan copyright jurisprudence going forward.
The finding that USSD structural elements constitute protected expression, without engagement with the software-specific authorities that address this question.
The calibration of damages to a defendant's overall revenue rather than to the wrong itself.
The imposition of a perpetual royalty as a compulsory licence without statutory anchor.
That is the precedent every Kenyan creator and every business in the country would then be working under.
Comments (0)