By Keren Weitzberg (11 January 2021)

‘I do not have any documentation that shows that I am a Kenyan,’ Adan lamented. ‘My children have a birth certificate that says: “single parent”. It’s like I had died. And yet I am alive.’

Adan is among tens of thousands of northern Kenyans who cannot obtain a national ID because their details appear on the government’s biometric refugee database. The Somali civil war brought hundreds of thousands of refugees into neighboring Kenya in the early 1990s. The influx of asylum seekers coincided with periodic droughts in northern Kenya, one of the country’s most politically and economically marginalized regions. In the intervening years, many ethnic Somali Kenyans slipped into the refugee system, falsely claiming to be escaping the civil war in order to access vital food aid, education, health care, or the coveted opportunity to relocate abroad. Now, because their fingerprints are captured on refugee databases, the very system responsible for protecting asylum seekers has effectively rendered them stateless.

Without an ID in Kenya, many basic economic and political rights are out of reach. One cannot move about freely, open a bank account, register a SIM card, access an M-Pesa mobile money wallet, enter into government and corporate offices, or gain formal employment. To prevent their children from experiencing a similar fate, Adan and his wife decided not to declare paternity on their birth certificates.

According to many civil society groups, Adan is a victim of double registration. The problem of double registration greatly accelerated after the Kenyan government, with support from the UNHCR, took over the refugee registration process in the country. This facilitated greater data consolidation and interoperability across the refugee and national systems. Nowadays, Kenya’s Automated Fingerprint Identification System (AFIS) includes the biometrics of all registered refugees, which can be run against the citizen registry. The National Registration Bureau is able to cross-check the fingerprints of those who apply for a national ID, effectively shutting the door of citizenship to anyone found on the refugee database.

Those caught between the refugee and citizenship systems have few modes of redress. This speaks to longstanding patterns of discrimination against northerners and Somalis, who are widely perceived to be a demographic and security threat within the country. Only amidst mounting pressure from northern MPs and civil society groups did the Kenyan government agree to vet those claiming to be citizens. Late last year, Adan was one of thousands of Kenyans who passed through this vetting process in the hopes of being deregistered from the refugee database. But in spite of years of government promises, Adan and many like him are still awaiting a resolution to their insecure, liminal status.

With funding from Privacy International and support from the civil society group Haki na Sheria, I have been tracking the ongoing dilemma of Kenyan citizens caught between these two databases—considered neither citizens nor refugees. As a member of this Datafication and Digital Rights in East Africa Network, I’ve shared and discussed this ongoing research with my fellow partners. Made up of academic researchers, civil society organisations and tech sector actors, this GCRF-funded network explores ways in which datafication in East Africa is reshaping the relations between citizens, states, and the wider market.

The problem of double registration in Kenya is yet another example of the dangers and risks of exclusion posed by centralized biometric systems, which have become increasingly ubiquitous in the humanitarian and development sectors in recent years. But it’s much more than just the latest case of techno-solutionism gone wrong.

Over the last decade, international actors across the humanitarian and development sectors, including the UNHCR, the UNDP, and the World Bank, have become vocal proponents of digital identity systems. Remarkably, technologies long associated with securitization, criminality, and surveillance (such as compulsory IDs and biometrics) have, in recent years, become constitutive of a politics of recognition and inclusion. Digital identity has transformed into “a solution for greater economic, social, and political inclusion” and, among its most quixotic proponents, a universal human right. Even scholars who are less than sanguine about these developments have suggested that biometric technologies and digital identity systems may open the door for new modes of redistribution and ‘post-social’ forms of care. The problems that have arisen in Kenya raise fundamental challenges for those who see in digital identity technologies the prospect for a ‘post-social’ politics of inclusion and redistribution.

Biometrics and exclusion

At the most basic level, the problem of double registration in Kenya serves as a cautionary tale about the rush towards centralized biometric systems and growing interoperability in the humanitarian sector. Scholars and civil society groups have long expressed concerns about the collateral damage caused by exclusion from ID initiatives. “Incessant proof of legitimate identity”, as Nikolas Rose argues, has become increasingly important for “conditional access to circuits of consumption and civility”, raising the stakes of exclusion. Kenyan citizens trapped in legal limbo have joined the ranks of marginalized populations worldwide who have been denied key government services due to automated biometric systems. The Kenyan ‘double registered’ are the victim both of longstanding, systemic discrimination as well as the unforeseen effects of new forms of datafication.

Data and institutional tensions within the humanitarian refugee sector

The problem of double registration in Kenya also exposes key institutional tensions at the heart of the refugee sector. Caught between host and donor countries, international bodies like the UNHCR must navigate a complex landscape of interstate sovereignty. The data-collection practices of humanitarian organizations often diverge or directly conflict with those of governments. While the UNHCR is primarily concerned with reducing fraud, host and donor countries are often most anxious about securitizing their borders. Host and receiving countries do not always share the UNHCR’s avowed commitment to humanitarian protection.

These entanglements result in contradictory, inconsistent views on data protection. Given Kenya’s poor record of refugee protection, UNHCR officials have long been hesitant to share information with Kenyan officials. Nevertheless, as part of an effort to shift greater responsibility onto host countries, the UNHCR funded and trained Kenya’s Department of Refugee Affairs (now the Refugee Affairs Secretariat). After much deliberation and under government pressure, the UNHCR also provided the Kenyan state partial access to its IT case management and biometric enrollment system. The agency’s centralized biometric infrastructure has enabled opaque data-sharing with third parties, such as the US Department of Homeland Security. Katja Jacobsen argues: “The very nature of digitalized refugee data means that it might also become accessible to other actors beyond the UNHCR’s own biometric identity management system.” Ironically, data collected for the purposes of humanitarian protection has now become grist to the mill of the Kenyan state bureaucracy.

This has also led to enormous ambiguities over data ownership and data responsibility. While the Kenyan government has ostensibly taken over responsibility for refugee registration and refugee status determination, the UNHCR continues to be involved in the day-to-day bureaucracy in myriad ways. Consequently, it’s unclear which government or international body is ultimately accountable for the undesired outcomes and unintended side effects of biometric data collection.

Fraud and the biometric ‘fix’

The UNHCR has painted biometric technologies as a fool-proof solution to the problem of fraud, an issue that has long troubled the agency. However, the concept of ‘fraud’ (and its pejorative connotations) often fails to capture the contingencies of refugee’s lived realities and the range of survival strategies adopted by those who skirt the boundary between citizen and refugee. Activities that one might label ‘fraud’ can also be understood as the outcome of national and global inequalities. With few legal options for travelling and relocating abroad, many people strategically engage with the refugee resettlement process. Others slip into the refugee system out of desperation. Accusations abound about refugees paying bribes to be resettled abroad. But this speaks as much to the corruption of UNHCR staff as it does to the fraudulent actions of refugees. Fraud is simply too blunt a term to contend with the moral ambiguities and moral imperatives of refugee identification.

As the problem of double registration in Kenya shows, there are merits to clunky, inefficient analog systems. The turn towards biometric and digital identity has not been as straightforwardly empowering as proponents often suggest. For decades, Kenyan Somalis could slip in and out of the refugee system with few consequences, taking advantage of much-needed aid and services from the UNHCR. The embrace of centralized, interoperable biometric systems may have helped in closing the data gaps between host countries, humanitarian organizations and donor states. However, deduplication efforts have also reduced the strategies available to populations who occupy such liminal positions. Datafication forced northerners into either/or categories, eliminating the informational voids through which many navigated precarity.

The assumed objectivity of biometrics has also trumped locally produced forms of identification. While many double registration victims have ample documentary evidence of citizenship (including birth certificates and school leaving certificates), the fingerprint has become the preeminent mode of ascertaining citizenship. As one of my interlocuters explained: “You can’t bypass the power of that fingerprint. It is a blockade.” Biometric scanners have become talisman-like objects, often used at the expense of other, potentially more reliable forms of narrative and documentary evidence. Datafication has shifted power in favor of a global technology embraced by the international humanitarian sector.

The Biometric Turn

It remains to be seen whether biometric and digital identity initiatives can be fully divorced from the gatekeeping mechanisms of the nation-state—including its exclusionary desire to police its borders and separate ‘insiders’ from ‘outsiders’. The UNHCR may have introduced biometrics to prevent citizens from slipping into the refugee system and refugees from registering multiple times. But the same techniques used to clean up the refugee database were then used to purge people from the citizenship registry.

As the case of Kenya shows, the introduction of centralized biometric systems increases the risk of function creep. Even if intended for humanitarian protection, data can be shared and technologies co-opted by donors, host states, and security agencies, leading to the further securitization of refugee policies. In Kenya, techniques intended to eliminate fraud and offer more efficient service delivery to refugees have become a means for the government to keep people out. However unintentionally, biometric data-sharing and data consolidation have barred some of Kenya’s most marginalized populations from accessing the benefits of citizenship, as though the refugee system were working in reverse.

This raises troubling questions for those who see biometrics as a potential technology of inclusion, capable of paving the way for a new politics of recognition and redistribution. It also reveals fundamental tensions around identification. To gain rights, one must be recognized. But for northerners in Kenya, being recognized on one database has led to expulsion from another. Biometric accuracy at the technical level has caused significant bureaucratic errors. As the case of Kenya shows, there must be more consideration for the ways that biometric registration alters the stakes for the most vulnerable, shifting the balance between inclusion and exclusion, visibility and invisibility.

Keren Weitzberg is a Lecturer (Teaching) at UCL History and a Visiting Research Fellow at the UCL Institute of Advanced Studies.