Machine-Readable Refugees

Hassan (not his real name; other details have also been changed) paused mid-story to take out his wallet and show me his ID card. Its edges were frayed. The grainy, black-and-white photo was of a gawky teenager. He ran his thumb over the words at the top: ‘Jamhuri ya Kenya/Republic of Kenya’. ‘Somehow,’ he said, ‘no one has found out that I am registered as a Kenyan.’

He was born in the Kenyan town of Mandera, on the country’s borders with Somalia and Ethiopia, and grew up with relatives who had escaped the Somali civil war in the early 1990s. When his aunt, who fled Mogadishu, applied for refugee resettlement through the United Nations High Commissioner for Refugees, she listed Hassan as one of her sons – a description which, if understood outside the confines of biological kinship, accurately reflected their relationship.

They were among the lucky few to pass through the competitive and labyrinthine resettlement process for Somalis and, in 2005, Hassan – by then a young adult – was relocated to Minnesota. It would be several years before US Citizenship and Immigration Services introduced DNA tests to assess the veracity of East African refugee petitions. The adoption of genetic testing by Denmark, France and the US, among others, has narrowed the ways in which family relationships can be defined, while giving the resettlement process the air of an impartial audit culture.

In recent years, biometrics (the application of statistical methods to biological data, such as fingerprints or DNA) have been hailed as a solution to the elusive problem of identity fraud. Many governments and international agencies, including the UNHCR, see biometric identifiers and centralised databases as ways to determine the authenticity of people’s claims to refugee and citizenship status, to ensure that no one is passing as someone or something they’re not. But biometrics can be a blunt instrument, while the term ‘fraud’ is too absolute to describe a situation like Hassan’s.

Biometrics infiltrated the humanitarian sector after 9/11. The US and EU were already building centralised fingerprint registries for the purposes of border control. But with the start of the War on Terror, biometric fever peaked, most evidently at the borders between nations, where the images of the terrorist and the migrant were blurred. A few weeks after the attacks, the UNHCR was advocating the collection and sharing of biometric data from refugees and asylum seekers. A year later, it was experimenting with iris scans along the Afghanistan/Pakistan frontier. On the insistence of the US, its top donor, the agency developed a standardised biometric enrolment system, now in use in more than fifty countries worldwide. By 2006, UNHCR agents were taking fingerprints in Kenya’s refugee camps, beginning with both index fingers and later expanding to all ten digits and both eyes.

Reeling from 9/11, the US and its allies saw biometrics as a way to root out the new faceless enemy. At the same time, for humanitarian workers on the ground, it was an apparently simple answer to an intractable problem: how to identify a ‘genuine’ refugee. Those claiming refugee status could be crossed-checked against a host country’s citizenship records. Officials could detect refugees who tried to register under more than one name in order to get additional aid. Biometric technologies were laden with promises: improved accountability, increased efficiency, greater objectivity, an end to the heavy-handed tactics of herding people around and keeping them under surveillance.

When refugees relinquish their fingerprints in return for aid, they don’t know how traces of themselves can travel through an invisible digital architecture. A centralised biometric infrastructure enables opaque, automated data-sharing with third parties. Human rights advocates worry about sensitive identifying information falling into thehands of governments or security agencies. According to a recent privacy-impact report, the UNHCR shares biometric data with the Department of Homeland Security when referring refugees for resettlement in the US. ‘The very nature of digitalised refugee data,’ as the political scientist Katja Jacobsen says, ‘means that it might also become accessible to other actors beyond the UNHCR’s own biometric identity management system.’

Navigating a complex landscape of interstate sovereignty, caught between host and donor countries, refugee aid organisations often hold contradictory, inconsistent views on data protection. UNHCR officials have long been hesitant about sharing information with the Kenyan state, for instance. Their reservations are grounded in concerns that ‘confidential asylum-seeker data could be used for non-protection-related purposes’. Kenya has a poor record of refugee protection. Its security forces have a history of harassing Somalis, whether refugees or Kenyan citizens, who are widely mistrusted as ‘foreigners’.

Such well-founded concerns did not deter the UNHCR from sharing data with, funding and training Kenya’s Department of Refugee Affairs (now the Refugee Affairs Secretariat), which since 2011 has slowly and unevenly taken over refugee registration in the country. The UNHCR hasconducted joint verification exercises with the Kenyan government to weed out cases of double registration. According to the anthropologist Claire Walkey, these efforts were ‘part of the externalisation of European asylum policy … and general burden shifting to the Global South’, where more than 80 per cent of the world’s refugees live. Biometrics collected for protection purposes have been used by the Kenyan government to keep people out. Tens of thousands of ethnic Somali Kenyan citizens who have tried to get a Kenyan national ID have been turned away in recent years because their fingerprints are in the state’s refugee database.

Over the last decade, biometrics have become part of the global development agenda, allegedly a panacea for a range of problems. One of the UN’s Sustainable Development Goals is to provide everyone with a legal identity by 2030. Governments, multinational tech companies and international bodies from the World Bank to the World Food Programme have been promoting the use of digital identity systems. Across the Global South, biometric identifiers are increasingly linked to voting, aid distribution, refugee management and financial services. Countries with some of the least robust privacy laws and most vulnerable populations are now laboratories for experimental tech.

Biometric identifiers promise to tie legal status directly to the body. They offer seductively easy solutions to the problems of administering large populations. But it is worth asking what (and who) gets lost when countries and international bodies turn to data-driven, automated solutions. Administrative failures, data gaps and clunky analogue systems had posed huge challenges for people at the mercy of dispassionate bureaucracies, but also provided others with room for manoeuvre.

Biometrics may close the gap between an ID and its holder, but it opens a gulf between streamlined bureaucracies and people’s messy lives, their constrained choices, their survival strategies, their hopes for a better future, none of which can be captured on a digital scanner or encoded into a database.