How a Hunger Crisis Turned Kenyan Citizens into Accidental Refugees — and Cost Them Their Identity
Kakuma, 30 May 2026
A sweeping investigation reveals how biometric systems, built to fight aid fraud, have stripped thousands of Kenyan citizens of their national identity. Caught between survival and bureaucracy, citizens who registered for famine relief in 2009 now find their fingerprints flagged as refugee records, leaving them legally stateless in their own country.
A Famine, a Walk, and a Database That Never Forgets
In May 2009, Farrah Budhul Uleh, a Kenyan Somali from Garissa County — a town located approximately 500 kilometres north-east of Nairobi — made a desperate decision [2]. With his 11-member family facing starvation during one of the worst droughts in recent Kenyan history, he travelled 176.2 kilometres from Balambala to the Dadaab refugee camp, not as a refugee, but as a starving Kenyan citizen seeking food [2]. The drought had by December 2008 been declared a national disaster, triggering a US$406 million aid appeal to feed an estimated 10 million people, with the government and the World Food Programme (WFP) jointly distributing relief food to 1.4 million Kenyans [2]. As the late President Mwai Kibaki warned at the time: “These people will not be able to meet their minimum food requirements between now and the end of August 2009 without emergency measures” [2].
How Biometrics Became the Architecture of Exclusion
The biometric infrastructure that ensnared Uleh was not built overnight. Its origins stretch back to the 1999–2001 Kosovo crisis, when Microsoft, Hewlett-Packard, and Compaq provided hardware and software that led the UNHCR to launch Project Profile and the Profile Global Registration System (proGres) in 2002 [1]. Swiss firm Elca Group subsequently developed proGres versions V1 to V3, which were deployed in Kenya in 2004 as a biographical and facial photo registration tool [1]. A 2019 UN financial audit would later find that proGres V3’s offline local databases failed data protection requirements, posing direct risks to refugee safety [1]. In 2012, the UNHCR launched a call for proposals to expand Project Profile using fingerprints, iris, and facial recognition, contracting Accenture to deploy the centralised Biometric Identity Management System (BIMS), which connects global verification stations to Geneva and operates in conjunction with the proGres database using Accenture’s Unique Identity Service Platform (UISP) and Biometric Matching Engine (BME) [1].
The April 2016 Handshake That Made Kenyans Stateless
The moment that most directly determined the fate of people like Uleh came in April 2016, when the UNHCR shared access to its proGres database with the Kenyan government [3]. From that point, Live Capture Units (LCUs) introduced at Huduma Centres began capturing biometric data and creating direct interconnectivity between the proGres v4 database and Kenya’s national registration system [3]. The consequence was immediate and devastating: Kenyan citizens whose fingerprints existed in the refugee database were rendered de facto stateless [3]. Kenya’s National Registration Bureau (NRB) utilises the Automated Identification System (AFIS) alongside the Integrated Population Registration System (IPRS) to verify biometric and biographical data — and when a young Kenyan from Garissa or Wajir turned 18 and applied for a national identity card, a fingerprint match with the refugee database could result in outright rejection [3].
Courts, Deadlines Missed, and a System Still Being Fixed
Legal intervention has provided some relief, but enforcement has been slow. On 21 January 2025, the High Court in Garissa declared the government’s refusal to deregister vetted Kenyans from the UNHCR proGres database unconstitutional [2]. However, Haki na Sheria reported in January 2026 that the government had missed the court’s 60-day vetting deadline [alert! ‘the specific date of the 60-day deadline is not stated in the source’]. Limited vetting exercises did eventually begin in Dadaab and Masalani in October 2025 [2]. Justice John Ongiego captured the human cost plainly in his ruling, describing victims as having been “forced to seek refugee status in their own country because of survival” [2]. Constitutional expert Bobby Mkangi was equally direct: “If they were original Kenyan citizens, their citizenship, hence IDs, can’t be revoked” [2].
Accountability to Donors, Not to Refugees
The human consequences documented by this investigation sit within a broader critique of how humanitarian technology has been governed. Professor Gianluca Iazzolino, an academic who has studied these deployments, put the structural problem bluntly: “It is always to increase accountability to donors. What is always missing is accountability towards refugees” [1]. Lawyer and migration rights specialist Petra Molnar echoed this concern, noting that humanitarian agencies “set the agenda in terms of prioritisation when it comes to humanitarian innovation and technological development [and] get to designate spaces of technological experimentation with little oversight or accountability” [3]. The technical record supports these concerns: scanner failures on dirty or damaged fingerprints have been documented, BIMS’s facial recognition matching was excluded by 2016 due to outdoor lighting challenges, and a 2002 iris recognition deployment may have wrongly denied as many as 11,800 claimants [1].