How Nairobi's Digital Records Became a Hall of Mirrors: The Story Behind the Duplicate Image Crisis

Kenya's national digital infrastructure has a clutter problem. Across government-linked databases, civic tech platforms, and the sprawling network of e-services that Nairobi's Silicon Savannah built its reputation on, duplicate images — the same photograph, ID scan, or document stored multiple times under different file names — have quietly ballooned into a storage and integrity crisis that administrators are only now beginning to confront head-on.

The timing matters. The Ruto administration's ongoing austerity programme, structured around IMF fiscal commitments, has forced every ministry to account for wasteful spending. Storage infrastructure is not cheap. When a government data centre in Upper Hill is paying for three or four copies of the same file, that cost compounds across millions of records. The duplicate image problem is, at its core, a money problem as much as a technical one.

How the Backlog Built Up

The roots go back to the mid-2010s. Kenya's push toward e-government accelerated sharply after the 2013 elections, when pressure mounted to move citizen services online. The Huduma Centre network — with flagship locations on Mama Ngina Street in the CBD and in Westlands — processed enormous volumes of documents. Scans were uploaded through multiple channels: front-desk terminals, back-office batch uploads, and later mobile submission tools. Nobody standardised the naming conventions. Nobody ran deduplication checks at the point of ingestion.

The problem compounded during the COVID-19 era. Between 2020 and 2022, Nairobi's tech community scrambled to digitise everything faster — land records, health files, business licences. Startups working on government contracts often used different image compression formats and different hash algorithms, meaning two technically identical files could look like separate records to the system. iHub Nairobi, the co-working and innovation hub on Ngong Road that has anchored the Silicon Savannah identity since 2010, hosted several developer convenings during that period where the lack of interoperability standards came up repeatedly — but without a central mandate, individual contractors kept doing things their own way.

The Nairobi Metropolitan Services commuter rail digitisation effort, which accelerated after 2021 to link ticketing and passenger data across the Syokimau and Ruiru corridors, added another layer. Passenger photo IDs ingested at different station kiosks created parallel records with no cross-referencing. A commuter who registered at Syokimau station and later re-registered at the Nairobi Railway Station terminus could exist twice in the system, their portrait stored separately with no automated flag.

The Cost of Ignoring the Problem

Estimates of exactly how much redundant storage Kenyan public systems carry vary depending on methodology. The Communications Authority of Kenya's 2024 annual report noted that the country's public sector data volumes had grown by roughly 34 percent year-on-year for three consecutive years — a pace that outstripped any corresponding investment in data hygiene tools. That growth rate, cited in the Authority's published figures, reflects a sector adding new records without retiring old ones.

Private sector actors have started addressing the problem on their own terms. Safaricom's M-Pesa platform, which processes millions of KYC image submissions annually, moved to perceptual hashing for deduplication as far back as 2019. Commercial banks regulated under the Central Bank of Kenya adopted similar practices under the 2021 Digital Credit Provider regulations. Government systems, operating on older procurement cycles and tighter budgets, have lagged behind.

The Gen Z tax revolt of 2024, which forced President Ruto to withdraw the Finance Bill and triggered a broader audit culture around public expenditure, has indirectly accelerated pressure on ICT managers to justify infrastructure costs line by line. Storage bills are now scrutinised in ways they were not three years ago.

For administrators inheriting these systems, the practical path forward involves three things: deploying automated deduplication tools that work retrospectively on existing archives, setting enforceable ingestion standards for any new data pipeline, and — most critically — assigning a named officer at each agency with responsibility for data quality. Without that last step, the technical fixes will be undone within a year. The crisis did not happen because Nairobi lacks technical talent. It happened because accountability for digital housekeeping was nobody's job.