Focus on the Core Conflict

In an increasingly digital world, the pervasive influence of algorithms on our daily lives is undeniable. From personalized recommendations to targeted advertisements, algorithms shape our experiences in subtle yet profound ways. However, their impact extends far beyond commercial interests, posing significant challenges to the very foundations of democracy. As Thomas Christiano, a professor in the Department of Philosophy at the University of Arizona, argues in his work, the phenomena of filtering, hypernudging, and microtargeting, driven by algorithmic communications, can undermine the deliberative potential and political equality essential for a healthy democracy.

The Algorithmic Threat to Democratic Principles

Christiano identifies three primary ways in which algorithms challenge democratic norms:

Polarization and Undermined Deliberation: Filtering algorithms, which curate information based on past engagement, can create echo chambers, exposing individuals only to content that confirms their existing beliefs. This can lead to increased polarization within the electorate, as diverse viewpoints are minimized, and the common ground for constructive deliberation shrinks.
Erosion of Epistemic Potential: The rapid spread of fake news facilitated by algorithms can severely undermine the “epistemic potential” of broad democratic participation. When citizens are presented with misinformation, their ability to make informed decisions is compromised, thus weakening the very purpose of collective decision-making.
Threat to Political Equality: This is Christiano’s central concern. He argues that algorithmic communications can deepen political inequality by creating a divide between those who possess the means and sophistication to navigate or resist algorithmic manipulation, and those who are more vulnerable to it.

Democracy and the Vulnerability to Manipulation

Christiano posits that a significant vulnerability to manipulation arises from the nature of political engagement in modern democracies. Most citizens operate on what he terms “low-information rationality.” With limited time, money, and energy to dedicate to politics, individuals often rely on shortcuts and heuristics to make decisions (Downs 1957). This inherent constraint makes them particularly susceptible to tactics like hypernudging (continuously reconfigured communications that subtly steer choices) and microtargeting (highly specific messages directed at narrow demographic segments). Just as a patient with limited medical knowledge might be easily swayed by a doctor’s recommendation, citizens with low political information can be manipulated when faced with complex political choices.

A core tenet of Christiano’s argument is that political equality – the principle that all individuals should have equal shares of effective political power – is deeply threatened (Christiano 2008). This power encompasses not only resources for participation, like voting rights, but also informational power. In a world where complete political information is unattainable for most, the unequal distribution of informational power becomes a significant source of injustice (Christiano and Braynen 2008).

Informational power has two key dimensions:

Ability to understand and seek intelligible information: This involves having the resources and skills to grasp political issues, understand how policies align with one’s interests and the common good, and assess the reliability of information sources. Without these cognitive conditions, political participation can be akin to having a car without knowing how to drive or where to go – a resource without the power to achieve one’s aims.
Power to disseminate information: This dimension highlights the financial and structural requirements for effective communication in the political sphere, including election campaigns, media broadcasts, and algorithmically driven communications. These activities are often heavily financed by wealthier segments of society (Sorauf 1992), whose interests may not always align with the broader population. For instance, higher-income groups tend to favor less redistribution than lower-income groups when informed (Erikson 2015). This creates a significant “conflict of interests and aims” between those disseminating information and many of its recipients, fulfilling a crucial condition for wrongful manipulation.

The Role of Social Conditions and Cognitive Dependence

Christiano argues that the uneven distribution of social conditions profoundly impacts an individual’s cognitive abilities in politics. We live in a society where the quality of education and access to diverse information networks vary drastically. This leads to vastly different levels of political sophistication among citizens.

A crucial concept here is cognitive dependence. In our complex society, individuals rely heavily on a “cognitive division of labor” (Hutchins 1995). We depend on experts—doctors, mechanics, financial consultants—for information we don’t possess ourselves. This reliance extends to political life, where citizens depend on others to be informed about crucial matters. The quality of the networks and environments in which this distributed cognition occurs directly impacts an individual’s ability to navigate political life effectively. When these networks are unreliable or biased, or when individuals lack the developed cognitive abilities to compensate, they become highly vulnerable to manipulation. This cognitive dependence is seen throughout the social sciences in the form of shortcuts, heuristics, and distributed cognition (Lupia 2017).

A Simple Model of Information Gathering and Vulnerability

Christiano illustrates this vulnerability with a two-group model, drawing on Downs’s (1957) economics of information. The model is based on the principle that individuals collect information until the marginal benefits equal the marginal costs. This leads to “rational ignorance” or low-information rationality, where people stop gathering information before being fully informed because the costs outweigh the perceived benefits.

The model introduces the concept of “free information“—information acquired without actively seeking it out for a specific purpose (e.g., through education, casual conversations, or work). Christiano contrasts two hypothetical groups:

Employers: Individuals whose work necessitates understanding law and politics (e.g., businesspersons, lawyers). They gain significant “free information” and develop political sophistication through their jobs and interactions within their professional and social networks. This high-quality free information equips them to seek further information discriminately and productively.
Employees: Individuals whose work is simpler, with less engagement with law or politics. They receive less high-quality free information from their work and social circles, often relying on less sophisticated sources like attention-grabbing media. They may also have received lower-quality education due to socioeconomic factors.

This disparity in “free information” creates a significant disadvantage for Employees. They have a weaker foundation from which to pursue further political information, making them more susceptible to messages that lack epistemic merit or merely confirm existing biases. Employers, conversely, due to their exposure to complex information and the need to discriminate sources, are less vulnerable to such manipulation. Notably, party identification is a weaker basis for choosing candidates than more extensive policy knowledge, and politicians are more responsive to those with such knowledge (Erikson 2015).

The hypothesis is clear: low-quality free information, coupled with a diminished ability to discriminate between messages, makes individuals highly susceptible to the kind of manipulation that is concerning. Algorithms significantly enhance this threat by efficiently identifying vulnerable individuals and delivering emotionally charged or misleading half-truths that exploit their cognitive weaknesses and existing biases. The ultimate danger lies in the capacity of these algorithms to promote specific ends, often undisclosed, by influencing those least equipped to resist.

Addressing Epistemic Inequality and Strengthening Democracy

Christiano concludes by proposing strategies to mitigate algorithmic manipulation and foster greater political equality. The core problem is the combination of unequal ability to finance algorithmic communications (often dominated by the wealthy with distinct interests) and the deeply unequal social conditions that impact citizens’ ability to critically assess information. This creates a situation where objective manipulation can readily occur, violating the fundamental principle of political equality.

To counter this, Christiano outlines three basic strategies:

Regulate Content and Platforms: This approach involves requiring platforms (like social media companies) to implement algorithms that ensure a diversity of opinion reaches users. This harks back to principles similar to the Fairness Doctrine in the United States, which was imposed on networks in the 1950s, ’60s, and ’70s and repealed in the late ’80s. A lesser remedy could involve platforms providing clear warnings when information circulated on the platforms is thought to be highly manipulative and dubious (Cohen and Fung 2021).
Strengthen Social Bases of Informational Power: This strategy focuses on empowering middle- and lower-class individuals to better respond to manipulative efforts (Christiano 2019). Christiano argues for enhancing the “cognitive conditions” necessary for political sophistication.
Empower Diverse Intermediate Institutions: This involves strengthening the financial and participatory power of various non-governmental organizations and associations, particularly those representing working people. This directly addresses the funding asymmetry in political communication and is a concern of election campaign finance law.

Christiano emphasizes a twofold approach: enabling individuals to acquire political sophistication and making the use of algorithmic communications more democratic. He suggests two main avenues for enhancing the epistemic conditions of less well-off persons:

Improve Education: A seemingly obvious yet often under-pursued solution in many countries, particularly the US, is to improve the quality of education in disadvantaged neighborhoods. A strong educational foundation provides the “free information” necessary for developing critical thinking skills.
Democratize the Workplace: Drawing on Downs’s conjecture (1957), Christiano argues that transforming the division of labor can significantly improve cognitive conditions. By empowering unions or introducing substantial worker participation in firms, ordinary workers would gain “free information” about the legal and political environment of their work. This not only enhances their understanding of their own interests but also fosters a broader capacity for discussion and debate, akin to democratic participation (Pateman 1970). Studies suggest that union members are better informed on economic and workplace issues than others in the same jobs (Kim and Margalit 2017; MacDonald 2019), and the usual tendency of politicians to be significantly more responsive to wealthier individuals (Bartels 2008) is bucked in representative districts with high union density (Becher and Stegmuller 2020).

Christiano asserts that individuals have a right to the cognitive conditions of effective participation (Christiano 2008). This is a positive right, grounded in their deep interest in political sophistication and effective participation. Unlike negative rights where individuals are largely responsible for securing the benefits, information has a unique property: one often doesn’t know what one doesn’t know. Therefore, individuals cannot be held solely responsible for lacking certain informational abilities, especially if their social and economic environments hinder their development. This implies a democratic justification for rights to participation in workplace affairs, either through democratic unions or democratically run workplaces.

Finally, with regard to the financing of communication and, in particular, algorithmically determined communication, Christiano advocates for enhancing the presence and power of working people’s associations such as unions (O’Neill and White 2018). These can counteract the tremendous advantage that wealthy groups have in sponsoring communications that attempt to persuade people in society, thereby undercutting the possibilities of manipulation arising from great asymmetries of informational power. Proposals for increasing union membership include providing greater protection for union organizing drives, given that the majority of US workers desire union membership (Freeman 2007). Furthermore, some have proposed systems for popular finance of a broad array of associations that can represent the interests of persons by distributing vouchers specifically for this purpose (Ackerman and Ayres 2002).

Conclusion

This article has aimed to articulate a conception of objective manipulation and its critical relevance to democratic thought, specifically illustrating how increasingly prevalent algorithmic communications play a role in this phenomenon. It has highlighted why manipulation poses a significant problem for democracy, explored its underlying causes, and proposed some conclusions regarding necessary institutional reforms.

All of this remains an area of active and evolving study, particularly concerning the intricate mechanisms of algorithmically determined communications. This paper is intended as a foundational framework for further reflection on the central issues of political equality presented by the rise of algorithms in public discourse.

References

Ackerman, Bruce, and Ian Ayres. 2002. Voting with Dollars: A New Paradigm for Campaign Finance. Yale University Press.
Bartels, Larry M. 2008. Unequal Democracy: The Political Economy of the New Gilded Age. Princeton University Press.
Becher, Michael, and Daniel Stegmüller. 2020. “Union Presence and Responsiveness to Public Opinion.” Political Science Research and Methods 8(3): 510–26.
Benkler, Yochai, Robert Faris, and Hal Roberts. 2018. Network Propaganda: Manipulation, Disinformation, and Radicalization in American Politics. Oxford University Press.
Bovens, Luc. 2008. “The Ethics of Nudging.” Behavioral Public Policy 2(1): 1–18.
Christiano, Thomas. 2008. The Constitution of Equality: Democratic Authority and Its Limits. Oxford University Press.
Christiano, Thomas. 2019. The Rule of the Many: Fundamental Issues in Democratic Theory. Westview Press.
Christiano, Thomas. 2022. “Algorithms, Manipulation, and Democracy.” Canadian Journal of Philosophy 52(1): 109–124. doi:10.1017/can.2021.29.
Christiano, Thomas, and Jeffrey Braynen. 2008. “Democracy and Equality.” In The Oxford Handbook of Political Economy, edited by Barry R. Weingast and Donald Wittman, 715–30. Oxford University Press.
Cohen, G. A., and Archon Fung. 2021. Online Manipulation and the Democratic Dilemma. Oxford University Press.
Coons, Christian, and Michael Weber. 2014. Manipulation: Theory and Practice. Oxford University Press.
Danaher, John. 2017. “The Threat of Algorithmic Manipulation: An Ethical Analysis.” Philosophy & Technology30(2): 173–91.
Downs, Anthony. 1957. An Economic Theory of Democracy. Harper & Row.
Erikson, Robert S. 2015. “The American Public’s Policy Mood: A Review and New Evidence.” Public Opinion Quarterly 79(1): 1–25.
Freeman, Richard B. 2007. “What Do Unions Do? The 2004 Nobel Lecture in Economic Sciences, Redux.” Journal of Labor Research 28(4): 641–55.
Hutchins, Edwin. 1995. Cognition in the Wild. MIT Press.
Kahneman, Daniel. 2011. Thinking, Fast and Slow. Farrar, Straus and Giroux.
Kim, Dong-wook, and Yotam Margalit. 2017. “Labor Unions and Policy Responsiveness in the United States: A State-Level Analysis.” State Politics & Policy Quarterly 17(1): 3–23.
Lanzing, Michel. 2019. “The Hypernudge: Big Data and the New Era of Nudging.” Philosophy & Technology32(1): 1–19.
Lupia, Arthur. 2017. Uninformed: Why People Seem to Know So Little About Politics and What We Can Do About It. Oxford University Press.
MacDonald, Jessica. 2019. “Union Membership and Political Information Acquisition.” Labour Economics 59: 1–13.
O’Neill, Martin, and Stuart White. 2018. “Working-Class Power in the Digital Age.” In The Oxford Handbook of the Welfare State, edited by Francis G. Castles, Stephan Leibfried, Jane Lewis, Herbert Obinger, and Christopher Pierson, 804–17. Oxford University Press.
Pateman, Carole. 1970. Participation and Democratic Theory. Cambridge University Press.
Sorauf, Frank J. 1992. Inside Campaign Finance: Myths and Realities. Yale University Press.
Sunstein, Cass R. 2016. The Ethics of Influence: Government in the Age of Behavioral Science. Cambridge University Press.
Thaler, Richard H., and Cass R. Sunstein. 2009. Nudge: Improving Decisions About Health, Wealth, and Happiness. Penguin Books.
Wood, Allen W. 2014. “Manipulation.” Proceedings of the Aristotelian Society 114(3): 281–99.
Yeung, Karen. 2017. “‘Hypernudge’: Big Data as a Tool for Behavioural Regulation.” Modern Law Review 80(4): 635–79.
Zittrain, Jonathan. 2014. “Digital Gerrymandering.” Harvard Law Review 127(6): 1341–55.