The Algorithmic State? Challenges to Democracy in an Era of Digitalization

Abstract

The Algorithmic State? Challenges to Democracy in an Era of Digitalization Jane Fountain (bio) Ireland 2030 will be a country awash in artificial intelligence, data analytics, machine learning, and other advanced computing applications whirring away in the background throughout the economy and government. While these emerging technologies offer tremendous benefits, they pose equally serious challenges. More broadly, the continuing development of the information society has been accompanied by increasing economic and social inequalities, an existential sense by some of being left behind, loss of community, and – in the wake of increasingly diverse populations – pernicious growth in systemic racism residing in biased data and analytical models that pervasively underpin decision-making. This essay seeks to illuminate some of these challenges as a means to accelerate advancements in the democratic uses of artificial intelligence to promote the public good and economic well-being. In Leviathan, Thomas Hobbes argued in the seventeenth century that to prevent anarchy – the state of nature which results in a ‘war of all against all’ – subjects should grant the sovereign immense powers and should give up the notion of self-government. He further argued that state–society relations, in effect, should be organized around fear as a means to preserve order.1 Fortunately, in many countries democracy has prevailed. But the state and democratic institutions face challenges in the current day related to digitalization and the power that artificial intelligence potentially gives to the state. Without safeguards, the state might well change notions of privacy and vastly increase its powers of surveillance. As increasing inequalities provide fodder for political unrest, it is tempting to some to strengthen the hand of the state. The twentieth-century political theorist Robert Dahl recognised that growing complexity and technological dimensions of policy problems increasingly require technical expertise in political and policy decision-making. But he argued that technical experts may not possess the moral [End Page 92] qualities required to serve the public good and lack the general knowledge that should inform multifaceted decision-making. He wrote in Democracy and Its Critics that in such cases, power delegated to technical experts might become alienated power or a form of guardianship by technical elites.2 It may seem ironic today that Dahl did not view digitalization in the early 1990s as a challenge to democracy but as a potential tool to safeguard it. In Controlling Nuclear Weapons: Democracy versus Guardianship, he sketched requirements for what he frankly termed ‘quasi-utopian proposals […] to adapt technology to an urgent democratic goal’. His objective, he wrote, was to ‘stimulate further thought, and even action, so that what is now within our reach may soon become within our grasp’. Specifically, he called for shaping digitalization to make information ‘about the political agenda […] easily […] accessible to all citizens’; to create ‘opportunities to all citizens to influence the informational agenda, and to participate in a relevant way in political discussions’; and to ‘provide a highly informed body of public opinion that […] is representative of the entire citizen body’.3 He was frank as well in admitting that public opinion is not always highly informed but held out hope that digital information and communications might foster more fully informed citizens, as well as broader and more inclusive participation and informed deliberation. Like many, Dahl did not foresee disinformation, bullying in social media, and the trivialization of discourse through the demise of newspapers and journalism. Given recent developments in the growth of artificial intelligence and algorithmic decision-making, this essay asks if Leviathan’s rise is inevitable or if Dahl’s hopeful vision for digitalization as a tool of democracy is still possible. I. Digitalization, artificial intelligence, and computational algorithms Digitalization encompasses the development, implementation, and use of a range of computer-based technologies and applications, including the overlapping areas of artificial intelligence (AI), computational algorithms, big data, internet of things, and cybersecurity. Benefits of digitalization include greater efficiency, cost savings, increased effectiveness (during the pandemic, for example, in contact tracing, monitoring, modelling, and targeting subpopulations to prevent disease), streamlining and customizing public services, greater data sharing, data integration, and data analytics used for solving highly complex but stable problems.4 [End Page 93] Key digitalization challenges of importance to the...