Every election leaves behind lessons that matter more than the result itself, whether from a court judgment, a political miscalculation, or a quiet technical experiment that gets far less attention than it deserves. The recent Ol Kalou parliamentary by-election may end up remembered less for who won than for what it revealed about the future of election management in Kenya.
A Citizen-Led Digital Experiment
Most coverage understandably focused on the contest itself, while far less attention went to People’s IEBC, a citizen-driven initiative that trained polling agents to photograph the official Form 35A immediately after counting, upload the images to a digital platform, and use artificial intelligence to extract, validate, and aggregate the results. By midnight on the day after voting, the initiative had tallied results from 93 of the constituency’s 144 polling stations and published independently verified figures before the official declaration.
What Ol Kalou Really Demonstrated
The initiative did not prove that citizens can replace the Independent Electoral and Boundaries Commission (IEBC); citizens have always been able to count votes at the polling station, and party agents do it every election. What Ol Kalou demonstrated was narrower and more consequential: a polling-station result can now become a publicly inspectable digital object, rather than a paper document seen only by officials and a handful of agents.
That shifts the debate. For years, Kenya’s election arguments have circled one question: can we trust the electoral commission? The more useful question now may be whether election evidence itself can be independently verified by multiple actors at once, without producing confusion or undermining constitutional authority, because trust is a psychological state, while verification is an institutional process, and a democracy grows sturdier when it depends less on the former and more on the latter.
That is what happened, in miniature, in Ol Kalou, where citizens no longer had to simply believe an announced outcome but could examine some of the underlying evidence for themselves. The experiment did not create a rival electoral commission; it added a new layer to a structure of checks that already includes polling officials, party agents, domestic observers, and international missions. Independent digital verification by citizens using AI tools is, in practical terms, a new entrant to that list, and one that barely existed five years ago.
Progress and Emerging Challenges
That is real progress, but it also exposes problems Kenya needs to work through before 2027, and most of them are not primarily technical.
The first is coverage. A pilot that verifies results from a fraction of polling stations in one constituency is one thing, but a national election involves tens of thousands of stations across difficult terrain and uneven security conditions, and partial coverage cannot substitute for a complete national picture. Treating a partial sample as though it were comprehensive is itself a source of misinformation, and scaling this system nationally is as much a logistics and financing problem as a technical one.
Also Read: The Political Cost of Unfinished Grief in Ruto’s Re-election Calculus
The second is authenticity. A photograph is only as trustworthy as its documented origin: how does a citizen viewing an uploaded Form 35A know it actually came from the polling station it claims to represent, and that nothing changed between capture and publication? AI can flag some manipulation, but no algorithm replaces a proper chain-of-custody procedure, and few people in this debate are discussing what that procedure should look like.
The third is cybersecurity. Any platform publishing election data becomes a target for hacking, coordinated disinformation, or denial-of-service attacks timed to the moment results matter most, which means protecting election infrastructure now means protecting not only government systems but the independent platforms that increasingly shape what the public believes about an election in real time.
The fourth is governance: who audits the optical-character-recognition software doing the extraction, who verifies that an algorithm update hasn’t introduced a systematic error, and who is legally responsible if a public tally turns out wrong? A platform can be well-intentioned and still lack answers to any of these, because good intentions are not an institutional safeguard.
The fifth, and hardest, is authority in a moment of disagreement. Picture a future election where IEBC, the parties, and two or three citizen platforms each publish a slightly different tally before the official declaration, not through bad faith but because of missing forms or delayed uploads, while social media amplifies every discrepancy within minutes. Which number should the public act on while waiting for the official result? That is a constitutional question, not a software one, and Kenya has no settled answer to it yet.
Towards a New Election Integrity Architecture
None of this amounts to a case against People’s IEBC; it’s closer to the opposite, since an initiative only exposes its era’s next set of problems once it has actually worked. The right response is not celebration or suspicion but engagement with what comes next, and what comes next is a shift in the conversation from election technology to election integrity architecture.
Also Read: Kenya Youth Unemployment: Why Relief Programs Fail and Real Development Is Needed
Technology alone doesn’t produce credible elections, and neither does legislation written in isolation; credible elections increasingly require law, cybersecurity practice, AI governance, evidentiary standards, public communication protocols, independent auditing, and civic education to function as one system rather than five separate conversations.
A Glimpse of 2027
The timing matters, because 2027 will likely be Kenya’s most technologically monitored election yet, with AI, digital forensics, satellite connectivity, and citizen reporting all playing larger roles than five years ago. The question is not whether these tools will shape the election; they will. The real question is whether Kenya sets the rules before political competition intensifies, or waits until a dispute forces the issue.
Ol Kalou was a parliamentary by-election in Nyandarua County, and it was also a small, early glimpse of what election integrity may look like in the age of artificial intelligence. The next disputed election in Kenya may not be decided by who counted the votes fastest; it may be decided by which digital evidence the country has already agreed to trust, and that is a conversation worth having well before the first ballot is cast in 2027.
This article was written by George Nyongesa, a lecturer of philosophy and logic at the University of Nairobi and Chuka University.
Follow our WhatsApp Channel and X Account for real-time news updates.
