Browse
Richard Bertinet is a chef who has lived in the UK since 1988.1 He runs a well-known and popular cookery school in Bath and has penned several award-winning recipe books. A significant portion of the UK’s population is made up of people like Richard – people who migrated from EU Member States and made the UK their home. There is still no exact, official count of how many EU citizens are resident in the UK by virtue of free movement rights, but we now know it to be more than four million.2 That group is embedded within communities across all walks of life. Some have been in the UK for decades, while others arrived more recently. Following the leave vote at the June 2016 Brexit referendum, the status of this group quickly became uncertain. Quite apart from negotiating the rules that would apply, there was the immense challenge of how the new rules would be administered fairly and effectively at the speed required by the Brexit process. In response to this challenge, the Home Office adopted a novel process, known as the EU Settlement Scheme, which included a combination of online applications, partially automated decision making, and cross-departmental data-sharing arrangements. For people like Richard, it was, in the words of then Home Secretary Amber Rudd MP, meant to be ‘as easy as setting up an online account at LK Bennett’.3 Many applications were processed quickly and successfully.
Robtel Neajai Pailey is a Liberian academic, activist, and author, currently based at the London School of Economics and Political Science.1 Since 2006, she has applied for and obtained a range of visas for the UK, including as a tourist, a student, and a skilled worker. Pailey made several of her applications from the US, where she is a permanent resident. The application process was costly and a bit intrusive, but on the whole she felt the experience was ‘relatively smooth’. When Pailey applied for a visa from Ghana in 2018, however, she bore significant additional costs and delay. Between the Home Office, the British High Commission in Ghana, and the local visa application centre, no one seemed to know the status of her application or the location of her passport. The delay forced her to cancel a different trip at substantial personal cost, and her request for a refund of the application fees was refused. She described the experience as, simply, ‘the absolute worst’.
In recent years, the United Kingdom's Home Office has started using automated systems to make immigration decisions. These systems promise faster, more accurate, and cheaper decision-making, but in practice they have exposed people to distress, disruption, and even deportation.
This book identifies a pattern of risky experimentation with automated systems in the Home Office. It analyses three recent case studies including: a voice recognition system used to detect fraud in English-language testing; an algorithm for identifying ‘risky’ visa applications; and automated decision-making in the EU Settlement Scheme.
The book argues that a precautionary approach is essential to ensure that society benefits from government automation without exposing individuals to unacceptable risks.
The Home Office – the main UK public authority responsible for immigration – is keenly interested in identifying ‘sham’ marriages which are designed to game the immigration system.1 Since at least 2015, the department has used an automated system to determine whether to investigate a proposed marriage.2 Marriage registrars across the country transmit details of proposed marriages to the system via ‘data feeds’. The system applies eight ‘risk factors’ to assess the risk that a couple’s marriage is a sham. These risk factors include the couple’s interactions before the registrar, ‘shared travel events’, and age difference. The system allocates couples either a ‘green’ rating, indicating that no investigation is warranted, or a ‘red’ rating, indicating that an investigation is warranted to identify possible ‘sham activity’. This algorithm processes a large number of marriages each year. In a 12-month period across 2019 and 2020, the Home Office received 16,600 notifications of marriages involving a non-European national, of which 1,299 were subsequently investigated.
The three systems we have explored in this book barely scratch the surface of automation in government immigration systems. They are systems which have, for various reasons and through various means, come into public view. But automated systems are being developed and deployed in many more corners of the immigration bureaucracy. The current trajectory, both in the UK and around the world, is toward increasingly automated immigration systems.
From the transitional and experimental phase that we are currently in, it is clear that automated immigration systems can bring benefits. For example, automation has allowed millions of people to get their status under the EU Settlement Scheme quicker than would have otherwise been possible, reducing delay and associated anxiety. These systems also seem to have some success in reducing decision-making costs. However, automated systems also pose clear and real risks of failure. These failures can occur, and have already occurred, at both individual and systemic levels, with disastrous effects for individuals and their families, as well as wider society and the economy. The resultant harms must be taken seriously, and certainly more seriously than the Home Office appears to have taken them to this point.
At dawn on 30 June 2014, Raja Noman Hussain awoke to find about 15 immigration and police officers raiding his house.1 Raja, a 22-year-old Pakistani man, had arrived in the UK several years earlier to study. Now he was being accused of cheating in an English language proficiency test approved by the Home Office, which he had sat in 2012 to meet a condition of his visa. After confirming his ID, the officers told him to grab some clothes, handcuffed him, and took him into immigration detention. Raja spent the next four months in detention, during which time he estimates he met over 100 other international students who had also been detained on the same basis. What followed was six years of legal battles over the cheating allegation, which disrupted his studies, estranged him from his family, and cost him around £30,000. Finally, in early 2021, Raja succeeded in clearing his name and confirming his right to be in the UK.
Raja was one of the tens of thousands of students whose visas were revoked or curtailed – and studies disrupted or ended – after the Home Office accused them of cheating in a government-approved English language test. This scandal eventually hit the headlines. The ensuing appeals and judicial reviews – which became known as the ‘ETS cases’ – have cost the government millions of pounds.2 What is less appreciated about this debacle is that much of it centred on a failed automated system: a voice recognition algorithm which the government used to identify suspected cheats. This chapter explores that side of the story.
Despite the volumes written on digital politics, and notwithstanding their depth and scope, quality and clarity of arguments and insights from digital scholarship, there do seem to be some matters that require attention. In this spirit Evelyn Ruppert, Engin Isin and Didlier Bigo propose a more subtle, nuanced appraisal of ‘data politics’. They propose that digital networks, or more precisely the data they produce, reconfigures ‘relationships between states and citizens’, thereby generating ‘new forms of power relations and politics at different and interconnected scales’ (2017, 1, 2). They contrast this to the similar, albeit different, forms of calculation that feature in and facilitate modern European state formation. This comparison is apt given that Andrew Feenberg notes that ‘technology is one of the major sources of public power in modern societies’ (2010, 10). The key difference between these sets of literatures, Ruppert, Isin and Bigo argue, is that the digital one has yet to pin down its ‘subjects’. They suggest that this identification effort can best be achieved by employing the post-structuralist tools bequeathed by Michel Foucault and Pierre Bourdieu. Ruppert, Isin and Bigo summarize their approach by stating that ‘Data does not happen through unstructured social practices but through structured and structuring fields in and through which various agents and their interests generate forms of expertise, interpretation, concepts, and methods that collectively function as fields of power and knowledge’ (Ruppert et al, 2017, 3).
As the US contends with issues of populism and de-democratization, this timely study considers the impacts of digital technologies on the country’s politics and society.
Timcke provides a Marxist analysis of the rise of digital media, social networks and technology giants like Amazon, Apple, Facebook and Microsoft. He looks at the impact of these new platforms and technologies on their users who have made them among the most valuable firms in the world.
Offering bold new thinking across data politics and digital and economic sociology, this is a powerful demonstration of how algorithms have come to shape everyday life and political legitimacy in the US and beyond.
Platforms have distributed propaganda that cultivated bigotry, all the while being prone to security breaches. When coupled with the looting of economic sectors like journalism, plus the installation of mass surveillance infrastructure which collaborates with state and corporate entities, the emerging image is of firms whose routine operations are wholly adjacent to broad-based democratic imperatives. Moreover, the centrality of privately owned platforms to American culture is indicative of the extent to which capital has gained control of public discourse. This algorithmic public sphere presents a general impediment to democratization in the US and elsewhere. But this is only the departure point for an analysis of class rule and unfreedom in American life.
More broadly, conditions for capital accumulation have never been more favourable. But the efficiency of this social logic is necessarily bound together with the dramatic acceleration of global social inequality and thus the beginnings of revolutionary demands from the many who have been excluded and for whom it has come at their expense. One looping effect of this deprivation and the contradictions upon which it rests is that an organic crisis emerged in the US.
American politics has recently passed catastrophic equilibrium. On Twitter, Donald Trump performs his authoritarianism by labelling the news media as ‘the enemy of the American people’ (Trump, 2017a). Views like these are not to be lightly dismissed. As Trump proclaims, ‘more than 90% of Fake News Media coverage of me is negative’, and so, for him, ‘Social Media [is] the only way to get the truth out’ (Trump, 2017b). This manoeuvre is but one in a series of coordinated efforts by the Trump administration to routinely delegitimize media organizations like MSNBC and CNN to assert that he is the only valid source of information.
This technique has been very effective. Consider how the New York Times published a story based upon an 18-month investigation into Trump’s taxes, which include tax fraud and financial losses throughout the 1980s of $1.17 billion (Barstow and Buettner, 2018). But while the reporters were later awarded a Pulitzer Prize for their journalism, the story effectively dropped from the news cycle.
