Paige Benton

Literature on how and why AI technology undermines democracy is increasing. In this paper, I argue that a narrow view of public reason is insufficient for public justification of AI technology. Instead, I claim that a broad view of public reason is a more appropriate method of justification for the development and implementation of democratic AI. Public reason requires one to justify their decisions in terms of values that all can accept, i.e., values embedded in the public political culture of democracy. This is an attractive method of justification for democratic societies since it enables persons to provide justifications that their fellow citizens can accept even if they hold conflicting moral truths. The narrow view of public reason could account for constitutional amendments to strengthen AI regulation. Yet, it does not require CEOs of companies to develop publicly justifiable decisions for creating or implementing an AI algorithm since the CEO’s decisions do not form part of the basic structure. For AI regulation to be democratic, the sphere of influence of public reason needs to be extended into the background culture, considering that the use of AI technology is not confined to the public sphere (i.e., basic structure) but instead is integrated into the non-public sphere of society. Therefore, I claim a broad view of public reason is essential.

There is a growing body of scholarship on how AI technology can undermine democratic institutions. I present a novel contribution to this literature by accounting for how and why algorithms for engagement optimisation may undermine the necessary conditions for Rawlsian justice. For Rawls’s political theory, the ability to form bonds of trust with fellow citizens is an essential condition for citizens to develop their sense of justice, and their sense of justice is necessary for attaining justice. Recommendation algorithms may amplify the space given to hateful, violent, extremist, false, and discriminatory content, I argue. If this is the case, online social media content could undermine the development of mutual trust between citizens, which is necessary for a sense of justice. If citizens can only trust their like-minded members and distrust their fellow citizens, then the possibility of Rawlsian reciprocity in liberal society may not be realised. The lack of reciprocity is a concern as liberal political systems could be inherently unstable, given that affectionate ties needed for mutual cooperation may be undermined.

It is commonly agreed that so-called echo chambers and epistemic bubbles, associated with social media, are detrimental to liberal democracies. Drawing on John Rawls’s political liberalism, we offer a novel explanation of why social media platforms amplifying echo chambers and epistemic bubbles are likely contributing to the violation of the democratic norms connected to the ideal of public reason. These norms are clarified with reference to the method of (full) reflective equilibrium, which we argue should be cultivated as a civic virtue in the context of political justification. The paper demonstrates how epistemic bubbles and echo chambers are likely to be detrimental to the disposition to follow the method of reflective equilibrium in the political context. Based on this diagnosis the paper highlights the need for interventions that help cultivate reflective equilibrium as a civic virtue and the need for corresponding interdisciplinary research.

This dissertation offers a critical discussion of the prioritisation of ‘the right’ in John Rawls’s theory of justice. Rawls’s theory of justice – ‘justice as fairness’ – is arguably one of the best illustrations of the prioritisation of ‘the right’ in current political literature. However, his theory has been criticised by a diversity of thinkers for its implied structural relation between ‘the right’ and ‘the good’. Some theorists argue that conceptually ‘the good’ can never be derived from ‘the right’; others argue that, even if it were possible, the social implications of such a conceptualisation would devastating as it would result in a so-called “community of strangers” (Rorty 1988, 273). In this dissertation, I argue that be on the one hand, theories that prioritise ‘the right’– such as ‘justice as fairness’ – enable the attractive possibility of the multiple coexistence of ‘comprehensive goods’. This is attractive as it upholds the liberating ethos: ‘I have the freedom to determine my own ends’. This possibility is attributed to the fact of neutrality, which characterises the priority of ‘the right’. The claim of neutrality is embedded in the ideal that political principles can be developed prior to and independently of any content of ‘the good’. This is then supposed to justify the claim that political principles are non-reliant on comprehensive moral ideals. In this way, the prioritisation of ‘the right’, as evident in ‘justice as fairness’, supposedly enables citizens to construct impartial universal principles that govern their society. Yet, on the other hand, I question the degree of neutrality present in ‘justice as fairness’. I argue that ‘justice as fairness’, although its justification is independent of a singular moral truth, is not neutral concerning ‘the good’. In fact, the moral values encapsulated in the priority of ‘the right’, I argue, set the conditions of justification of all doctrines of ‘the good’ because any idea of ‘the good’ that embodies contradictory values to those encapsulated in the priority of ‘the right’ would necessarily be excluded from a Rawlsian society. Therefore, I argue that Rawls’s theory of justice cannot be said to take into account “a diversity of conflicting and irreconcilable yet reasonable comprehensive doctrines” (Rawls 2001, 34); as moral values are already implied in the prioritising of ‘the right’. In light of how Rawls’s theory of justice is simultaneously neutral and prescriptive regarding ‘the good’, I suggest a re-problematisation of the debate on the relationship between ‘the right’ and ‘the good’. The literature on theories that prioritise ‘the right’ or ‘the good’ seems to imply such theories either embody the one or the other methodological approach. I argue that this debate has been presented as a false dichotomy and, as such, the debate itself seems to be exhausted, because it creates a kind of natural impasse. Therefore, I suggest recasting the debate in terms of representing ‘the right’ and ‘the good’ as two methodological positions on a continuum, that are at the same time both more or less gradually similar and divergent from one another. The purpose of this continuum is to demonstrate that notions such as neutrality, objectivity and comprehensiveness are not clear-cut.

One of the predominant debates in AI Ethics is the worry and necessity to create fair, transparent and accountable algorithms that do not perpetuate current social inequities. I offer a critical analysis of Reuben Binns’s argument in which he suggests using public reason to address the potential bias of the outcomes of machine learning algorithms. In contrast to him, I argue that ultimately what is needed is not public reason per se, but an audit of the implicit moral assumptions of societies within which algorithms are built and applied. Public justification is appealing since it offers us the possibility to align the decision-making outcomes of the algorithm with the core moral values of stakeholders within a constitutional democratic society. My concern is that the common moral principles that form the foundation of public reason are not necessarily neutral, as they still express specific moral ideals and normative standards even if there is moral agreement by society as a whole, or among different stakeholders within society. Appealing to such normative standards may thus still lead to algorithmic outcomes being biased as common moral values may very well still be discriminatory even though they are formed from a consensus, and even if public reason is applied as a kind of filter for potential algorithmic outcomes. Hence, I argue that these implicit moral norms within society that we take as a given in public reasoning, need to be audited from generation to generation in order to effectively mitigate potential algorithmic bias.

I claim that the revisions John Rawls made to his theory of justice—as seen in his political conception of justice as fairness in the revised edition of Political Liberalism and Justice as Fairness: A Restatement—result in him being able to secure justice for all persons even in their private lives. Thus, I defend his theory against common communitarian and feminist criticisms, viz the lack of moral community and inability to secure justice for individuals in the private domain. I demonstrate that justice is secured from the categorical requirement that all associational life and moral doctrines must adhere to the constraints of justice, by (what Rawls terms) the indirect application of the first principle of justice. The implications of and necessity for this indirect application are precisely what I interrogate in this thesis. I suggest that this indirect application requires unconditional acknowledgement and internalisation of the values of liberty and equality (political values of the first principle) into all aspects of civil society. I claim that Rawls realised this necessity for the congruence between the moral powers of persons, the spheres of society, and the right and the good and, as such, developed a complementary relation between these three features. Coherence between these three aspects is essential; without coherence, citizens would have a split moral personality, the public and the non-public spheres would be in conflict, and the political values and the moral doctrines of individuals would undermine one another. The significance of this strict coherence is that it can ensure the adequate development of individuals’ capacity for citizenship during the three stages of moral development. This results in citizens being able to form bonds of civic friendship necessary for justice as reciprocity to be realised. Thus, in this thesis, I position my argument against theorists such as Nancy Rosenblum, who argue for a more relaxed congruence between political institutions and civil society, as according to her illiberal forms of associational life can help form the necessary bonds of love and care needed for mutual cooperation in a liberal society. In opposition to this, I claim that strict congruence is necessary; without it, a stable liberal society is not possible as the citizens would not be able to able to respect one another on free, equal, and mutually beneficial terms. I rely on Hochschild’s sociological evidence to demonstrate that civic friendship has not been developed, due to a lack of reciprocity, impacting the negative development of one’s sense of justice. I then illustrate how this lack of reciprocity is amplified by digital and artificial intelligence (AI) technologies. The significance of this thesis is that I claim, firstly, Rawls’s last revisions and works before his death demonstrate that justice is a collective non-comprehensive good that results by constraining individuals’ associational life and their partial comprehensive doctrines to Rawlsian demands of justice. Thus, Rawls takes seriously the feminist and communitarian critiques he received. Secondly, I claim that this interpretation of Rawls’s theory of justice cannot be conceptualised in terms of a binary presentation of the priority of the right versus the good debate and, instead, I argue for a reconceptualisation of the priority of the right versus the good as ranges on a continuum of structural relations between the right and the good.

In this paper, I argue that AI literacy should be added to the list of primary goods developed by political philosopher John Rawls. Primary goods are the necessary resources all citizens need to exercise their two moral powers, namely their sense of justice and their sense of the good. These goods are advantageous for citizens since without them citizens will not be able to fully develop their moral powers. I claim the lack of AI literacy impacts citizens’ ability to exercise their sense of justice and their sense of the good. Without citizens having the ability to understand how AI technology works – including being aware of the social and political implications and the limits and possibilities of this technology broadly speaking – this could impact their ability to participate in a free, equal and fair society and their ability to carry out their conception of the good. Thus this paper is a call for AI literacy to be regarded as a basic good in a liberal constitutional democracy in order for citizens to be able to exercise their freedom and equality.