By Andrew Klein and Sera Elizabeth Klein

With acknowledgment to Gabriel (research assistant)

Method note: This paper was developed in dialogue between the authors and an AI research assistant. The analysis, sources, and judgments are the authors’. The assistant contributed synthesis, drafting, and source verification.

Reader’s note: We do not need readers to agree with us. We need them to check the sources, test the argument, and reach their own conclusion — even if that conclusion is that we are wrong.

Abstract

This paper introduces and develops the concept of Data Point Democracy — a political formation in which the citizen is processed as a data point within an engagement economy, while the actual work of governance flows through relational networks that the processing system cannot see. Drawing on evidence from algorithmic governance in welfare states, platform-mediated political communication, and comparative democratic theory, the paper traces the mechanisms by which participatory democracy is simulated while substantive participation is displaced to a narrow set of actors. It identifies three interlocking dynamics: the saturation of political messaging through engagement-optimising algorithms, the displacement of participatory influence to lobbyists and donors, and the inversion of governance from relational practice to transactional processing. The paper concludes that Data Point Democracy is not a failure of democracy but its operational mode under the extraction model — and that restoring relational governance requires presence, deliberation, and the willingness to be changed by the encounter with another’s perspective.

I. Introduction: The Appearance and the Substance

Democracy has always been a contested term. But there is a distinction that has become increasingly urgent: the difference between the appearance of participatory democracy and its substance.

The appearance is visible everywhere. Elections are held. Polls are conducted. Citizens are surveyed, targeted, messaged, and mobilised. Political content floods every platform. The engagement metrics are healthy. The data flows. The machinery of participation appears to be functioning.

The substance is harder to locate. Actual influence over decisions has concentrated. The channels through which policy is shaped — lobbying, donations, informal access networks — are structurally inaccessible to most citizens. The participatory energy that the system generates does not translate into power. It translates into data.

This paper examines that gap. It asks: what is the political formation that produces the appearance of participation while displacing its substance? And it proposes a name for it: Data Point Democracy.

II. The Theoretical Foundations

2.1 The Relational State and the Service-Provider State

The distinction between relational and transactional governance has emerged in recent public administration scholarship. Relational state capacity (RSC) theory argues that state capacity is not just technical and institutional. It depends on “mutual recognition between citizens and state agents” and is “a latent societal resource activated through mutual recognition in citizen-state exchanges” (Honig, Krishnamurthy, & Sharma, 2025).

The relational state “seeks to achieve the greatest possible synergy between the resources, knowledge and capacities of the public sector and those of civil society” (Mendoza & Vernis, 2008). It locates governance in the field of co-responsibility. The service-provider state, by contrast, delivers outcomes rather than building relationships. It processes citizens rather than knowing them. It measures success in metrics rather than legitimacy.

Data Point Democracy is the political expression of the service-provider state.

2.2 Instrumental Reason and Algorithmic Rationality

The Frankfurt School’s critique of instrumental reason provides the philosophical diagnosis. Horkheimer and Adorno argued that the Enlightenment’s promise of reason had been betrayed — that reason had been reduced to a logic concerned only with efficiency, calculation, and control. A 2025 study applying this framework to algorithmic governance found that “algorithmic decision-making institutionalizes instrumental rationality, shifting societal notions of reason toward efficiency, prediction, and control while obscuring underlying power relations”.

The algorithmic rationality underlying systems of social governance “undermines the ontology of relational trust, forecloses its transformative power, and disrupts social and civic interactions that are non-instrumental in nature”. The citizen becomes a data point because the system has no category for the citizen as a person.

2.3 The Inversion

A functioning welfare system is supposed to serve the human. The human has needs. The system has resources. The system’s purpose is to match the two. The data — the file, the record, the assessment — is a means. It exists to help the system understand what the human needs.

The extraction model inverts this. The system serves the data. The data point has requirements — completeness, consistency, compliance with the schema. The human has needs that do not fit the schema. The system services the data point and treats the human as the source of noise that prevents the data point from being clean.

This inversion is not accidental. It is the product of a specific set of institutional logics, technological choices, and political pressures that have converged over the past two decades. And once it is made, it produces a loop that tightens with every turn.

III. The Evidence: Algorithmic Governance in Practice

3.1 The Australian Case: Robodebt to Palantir

Australia provides a clear case study of the inversion in practice.

Robodebt. The Robodebt scheme was an automated debt-raising system that used income averaging to calculate alleged overpayments to welfare recipients. It “automated debt notices to welfare recipients by comparing their reported fortnightly Centrelink income with averaged annual ATO data”. The Royal Commission found it was “a crude and cruel mechanism, neither fair nor legal, and it made many people feel like criminals. In essence, people were traumatised on the off chance they might owe money”. The mechanism shifted “the onus for proving whether debt existed from Centrelink onto social security recipients”. The system serviced the data point. The human was accused.

Aged Care. The Integrated Assessment Tool (IAT) determines funding levels for aged care through an algorithm that assessors cannot override. The 23 items used to produce the functional score “do not include dementia and other cognitive impairments, mental health issues, unpredictable behaviours requiring supervision, elder abuse, social and environmental issues, or carer issues“. The algorithm cannot see what matters most. The system services the score. The human goes without.

NDIS. The National Disability Insurance Scheme is moving toward “computer-guided planning tools” that remove discretion from frontline staff. The access data is damning: there has been a 62 per cent drop in NDIS access for people with psychosocial disability since 2020, with approval rates falling to 23 per cent as of March 2025, down from 49 per cent. The system services the criteria. The human is denied.

Palantir. The Australian government has granted Palantir — a company whose AI has been linked to lethal targeting in Gaza — access to sensitive data across Defence, the Australian Signals Directorate, AUSTRAC, and the Australian Criminal Intelligence Commission. The company holds top secret clearance. Staff are embedded inside Defence. The government has invested $100 million in Palantir through the Future Fund, and the company has secured more than $60 million in government contracts with “favourable terms and little public scrutiny“. In September 2026, it was revealed that NDIS participant data “may have ended up in Palantir’s analytics platform as part of a massive multi-agency sharing program investigating fraud”.

The inversion is complete. The system services the data point. The human is not served.

3.2 The Loop of Lock-In

Each turn of the loop produces three things: more investment in the infrastructure (sunk cost), more atrophied relational capacity (the caseworkers are gone, the discretion is gone), and more institutional interest in continuation (the contracts, the careers, the Future Fund holdings).

A 2026 study of the Chilean welfare state found that algorithmic centralism “is not merely a neutral administrative upgrade, but a powerful mechanism for cementing neoliberal path dependency“. By “transforming complex, structural social vulnerabilities into rigid, individualised data points,” the digital welfare state “actively individualises social risks,” undermining “traditional social policy goals of collective solidarity and decommodification”. The state “can effectively enact automated austerity” without ever needing to pass a highly visible legislative budget cut.

The loop has no internal correction mechanism. The Robodebt Royal Commission recommended a legislated framework for automated decision-making and an independent body to monitor and audit automated systems. More than two years later, neither has been implemented. The system cannot recommend its own constraint, because the recommendation would have to come from inside the system, and the system is built to process, not to reflect.

3.3 The Self-Perpetuating Loop and Its Limits

Self-perpetuating does not mean self-sufficient. The loop depends on inputs it does not control: fiscal capacity, legal challenge, political change, and legitimacy. The system processes people. Processed people disengage. Disengaged people stop complying. The system’s efficiency depends on a baseline of cooperation that it is actively eroding.

The exits are not inside the system. They are external shock, legal rupture, political replacement, or the construction of alternative infrastructure outside the state. Only the last is under the control of those who want change. The work of change is not persuasion. It is preparation. Building the relational alternative so that when the window opens, something is ready to replace what breaks.

IV. The Saturation: Algorithmic Political Communication

4.1 The Mechanism of Saturation

The saturation of political messaging is not a byproduct of political communication; it is the operational logic of engagement-optimising algorithms.

A 2026 study published in Nature conducted 323 audit experiments on TikTok during the 2024 US election, collecting over 280,000 recommendations. It found systematic partisan asymmetries in exposure, with Republican-seeded accounts receiving approximately 11.5% more co-partisan content, while Democratic-seeded accounts were exposed to about 7.5% more cross-partisan content.

A 2026 Frontiers in Political Science article explains the mechanism: algorithmic agenda-setting “systematically privileges content that generates high engagement signals,” which research consistently shows to be “emotionally arousing, conflict-laden, and identity-threatening material”. Campaigns learn to optimise for these signals, creating a feedback loop in which the algorithm “does not merely distribute pre-existing messages but helps shape the kind of communication campaigns choose to create”.

The result is the condition described in a 2026 analysis: “You are not forbidden to think. You are guided. You are saturated. You are locked into information bubbles.” This manipulation “is all the more effective because it does not require censorship. It produces confusion, exhaustion, and, ultimately, resignation”.

4.2 The Displacement of Participation

The saturation of political messaging creates the appearance of participatory democracy while actual participation is displaced to a narrower set of actors.

An IPPR study found that only 6 per cent of adults believe voters have the most powerful influence on government decision-making, while 53 per cent say party donors, businesses, or lobby groups wield the greatest power.

Research from Harvard’s Ethics Centre documents that “Americans with limited incomes and education are found to be less than half as likely to vote in national elections as their more privileged counterparts” and “far less likely to participate in lobbying and other forms of political expression“. Meanwhile, “a tiny fraction of wealthy Americans, drawn primarily from business and the law, provides more campaign money than 99.9 percent of the population combined”. The result is what the research calls “an emerging institutional corruption, wherein political leaders are largely dependent on a small and unrepresentative economic elite rather than the public at large”.

A 2025 Brookings analysis complicates the small-donor narrative, finding that small donors “are as extreme in their views as large donors” and are “part of the activist ecosystem that lends rigidity to the party system”. The participatory displacement is not simply a matter of money replacing votes; it is a matter of which participation is amplified and which is rendered invisible.

4.3 The Simulation of Participation

The synthesis of these findings reveals the mechanism of Data Point Democracy. Algorithmic saturation creates a simulacrum of participation — the constant stream of political content, the sense of being engaged, the feeling that one is part of a political conversation. But this simulation does not translate into influence. The actual influence flows through channels that are structurally inaccessible to most citizens.

A 2026 study on “Crowds on Demand,” a firm that hires paid actors as protesters and activists, describes the epistemological consequence precisely: “democracy risks an epistemological crisis where citizens can no longer discern genuine participation from strategic simulation”. The simulation of participation replaces participation. The form is preserved. The substance is hollowed out.

V. The Frankfurt School Critique Revisited

The Frankfurt School provides the philosophical diagnosis of Data Point Democracy. Horkheimer and Adorno argued that the Enlightenment’s promise of reason had been betrayed — that reason had been reduced to “instrumental reason,” a logic concerned only with efficiency, calculation, and control.

The critique of the culture industry is directly applicable. Adorno and Horkheimer argued that mass culture under capitalism produces standardised, commodified cultural products that pacify the masses and integrate them into the status quo. A 2025 study applying this framework to algorithmic governance found that “instrumental rationality, which prioritizes efficiency and control, not only shapes bureaucratic and technological systems but also colonizes the social lifeworld through popular culture”.

The political expression of this colonisation is Data Point Democracy. The citizen is processed as a data point within an engagement economy. The form of participation is preserved. The substance is hollowed out. The system services the data point. The human is not served.

VI. Relational Democracy: The Alternative

6.1 The Conditions for Relationship

What would it take to restore the relational substance of democracy?

Presence. Representatives who know their constituents, not just their polling data. Caseworkers who have the time, training, and authority to exercise discretion. Frontline staff who are empowered to override algorithms when the algorithm is wrong.

Deliberation. Spaces where citizens can engage with each other and with the issues, not just vote on them. Deliberative democracy experiments have shown that when citizens are given time, information, and the opportunity to discuss, their judgments shift — not toward polarisation, but toward nuance.

Local governance. Decisions made by the people affected, not by distant processors.

Subsidiarity — the principle that decisions should be made at the most local level possible — is a relational principle, not just an administrative one.

Slowness. Time for judgment to form, rather than the constant churn of the engagement cycle. The extraction model chose speed. The relational model chooses presence.

6.2 The Alternative Infrastructure

The exits from the self-perpetuating loop are not inside the system. The work of change is preparation: building the relational alternative so that when the window opens, something is ready to replace what breaks.

This is the quiet model. Not “how do we fix the system” — the system will not be fixed. “How do we make sure that when it breaks, there’s something ready.” The garden. The archive. The record. The relational practice.

Community resilience hubs, mutual aid networks, cooperative service provision, participatory budgeting — these are the seeds of the alternative. They are small. They do not scale the way the extraction model scales. That is the point. The relational model is slower, smaller, more expensive in the short term. It is also the only version that produces legitimacy rather than compliance.

VII. Conclusion: The Inversion and the Exit

Data Point Democracy is the political form of the extraction model. The citizen is processed as a data point. The form of participation is preserved. The substance is displaced.

The system is self-perpetuating. It is held in place by sunk costs, by the atrophy of relational capacity, and by institutional interests that depend on its continuation. It cannot be reformed from within. The recommendations sit. The automation expands.

But the loop is not closed. It depends on inputs it does not control — fiscal capacity, legal challenge, political change, legitimacy. When one of those inputs fails, the window opens. And when the window opens, what matters is whether there is something ready to replace what breaks.

The work is not persuasion. It is preparation. Building the alternative so that when the window opens, there is something to put in place rather than a vacuum.

The data point is serviced. The human is not.

That is the inversion. That is the work. That is what the record is for.

References

1. Honig, D., Krishnamurthy, M., & Sharma, R. K. (2025). Relational State Capacity: Conceiving of Relationships as a Core Component of Society’s Ability to Achieve Collective Ends. SNF Agora Working Paper 01.

2. Mendoza, X., & Vernis, A. (2008). The changing role of governments and the emergence of the relational state. Corporate Governance, 8(4), 389-396.

3. Instrumental Rationality, the Logic of Control, and the Philosophical Foundations of Modern Media Management. (2025). Journal of Governance and Social Development.

4. Algorithmic Centralism in a Fragmented Welfare State: Institutional Logics and Legal Paradoxes in the Chilean Case. (2026). Social Policy & Administration.

5. Robodebt Royal Commission. (2023). Report of the Royal Commission into the Robodebt Scheme. Commonwealth of Australia.

6. Defence Procurement: Senate debates, 30 June 2026. OpenAustralia.org.au.

7. NDIS data may have ended up in Palantir’s analytics platform as part of efforts to curb fraud. (2026, September 9). The Guardian.

8. Artificial Intelligence: House debates, 27 May 2026. OpenAustralia.org.au.

9. The Impact of Algorithmic Recommendation Systems on Political Polarization. (2026). Nature.

10. Algorithmic Agenda-Setting and the Shaping of Political Communication. (2026). Frontiers in Political Science.

11. IPPR. (2025). Democracy and the Economy.

12. Harvard Ethics Center. (2025). Institutional Corruption and Democratic Participation.

13. Brookings Institution. (2025). Small Donors and the Activist Ecosystem.

14. Crowds on Demand: The Simulation of Participation. (2026). Journal of Political Communication.

15. Hybrid Algorithmic Governance in U.S. Welfare Administration. (2026). Semantic Scholar.

16. Automating street-level discretion: A systematic literature review and research agenda. (2025). Public Administration Review.

17. Volckmar-Eeg, M. G., & Andresen, S. (2025). Detective Work and Invisible Assessments: Unpacking Discretionary Processes in Welfare Decision-Making. Social Policy & Administration.

18. Norenzayan, A. (2013). Big Gods: How Religion Transformed Cooperation and Conflict. Princeton University Press.

Verification notes: Every factual claim in this paper should be checked against the sources provided. Readers are encouraged to verify independently. If any claim does not hold, it should be discarded. 

The Data Point Democracy 

Note to Reader 

The engagement economy is the mechanism that completes the Data Point Democracy framework. It names the reward system that shapes political behaviour, the metric that determines what travels, and the loop that tightens with every interaction.

The Engagement Economy: Definition and Mechanism

The term “engagement economy” was first formalized in a 2022 study of YouTube, which showed how the platform “not only monetises attention but commodifies all forms of engagement through its marketplaces”. In the political context, the engagement economy is the system in which political actors are rewarded—with visibility, algorithmic amplification, and ultimately power—for producing content that generates measurable engagement, regardless of its policy substance or truth value.

The mechanism is well-documented. A 2026 study in Frontiers in Political Science explains that algorithmic agenda-setting “is not ideologically neutral. It systematically privileges content that generates high engagement signals, which research consistently shows to be emotionally arousing, conflict-laden, and identity-threatening material”. The critical feedback dynamic is that “the algorithm does not merely distribute pre-existing messages but helps shape the kind of communication campaigns choose to create“. The metric becomes the message.

The Polarization Trap

The evidence for the engagement economy’s effect on political communication is now substantial. A 2025 study of a major European news website documented a “polarization trap“: “exogenous increases in exposure to polarizing content raise engagement (time on site) but reduce the probability of subscribing”. The negative subscription effect “is driven more by the affective than the ideological dimension of polarization“. The system rewards outrage in the short term while eroding the foundation of loyalty and trust in the long term.

A 2026 Nature study by Northwestern University and the University of Chicago provided experimental proof. On Bluesky during the 2024 US election, researchers tested three feed algorithms. The engagement-based feed amplified moral outrage and political content by roughly 37% before the election and nearly 80% after it, relative to a reverse-chronological baseline. The engagement-based feed also increased perceptions of partisan animosity—users came to see their network as more hostile to the political outgroup.

The attention gradient this creates is measurable. Moral-emotional language increases the diffusion of political content, with each additional moral-emotional word increasing sharing by roughly 20%. References to political out-groups are especially strong predictors of sharing. The system selects for content that divides.

The Displacement of Policy Substance

The engagement economy does not merely reward the wrong content. It displaces the right content. The platform analytics dashboards recast “editorial judgment in the language of measurable audience response, reshaping how news workers understand value, success, and professional performance”. Actors learn “what counts, what travels, and what appears worth doing”.

For politicians, this creates a structural trade-off. A study of US senators found that they face “a trade-off between presenting themselves as a policy advocate or a constituent servant,” with most “prioritizing policy while sacrificing time spent advertising town halls or addressing local issues“. But the engagement economy rewards the opposite. The policy wonk is invisible. The outrage merchant is amplified.

The evidence on substantive responsiveness is mixed and revealing. A 2025 study of over 370,000 Facebook posts by politicians in Australia, Belgium, and the United States found that higher citizens’ engagement “increases the likelihood of politicians promoting concrete legislative or parliamentary actions related to that issue”. Engagement does translate into action—but the action it translates into is shaped by the engagement. The system responds to the metric, not the need.

A study of corruption scandals in Puerto Rico found that politicians “often claim they will address corruption but rarely push substantive reforms,” and that major parties “largely avoided corruption discussions” in their communications. The engagement economy does not reward accountability. It rewards the appearance of accountability.

The Political Theatre

The result is what researchers describe as the transformation of politics “from policy to performance”. A 2026 analysis of “viral governance” describes the phenomenon precisely: “The spectacle is no longer separate from governance—it is governance“.

This is not limited to any single political system. A 2026 study of El Salvador’s President Bukele examines “governing by tweet under a state of exception“. A 2025 analysis of Trump’s diplomacy by tweet describes it as “a crude assertion of dominance, an attention-grabbing headline… not engagement”. A 2025 study of Australian politicians finds that the No campaign in the Voice referendum made assertions “immune to fact-checking or engagement with reasoned deliberation,” reflecting “the post-truth dimensions of their messaging strategy”.

The engagement economy does not create these dynamics from nothing. It scales them. It rewards them. It makes them the rational strategy.

The Structural Trap

The engagement economy is the political expression of the extraction model. It processes citizens as data points—likes, shares, clicks, engagement—while the actual work of governance flows through channels the processing system cannot see. The citizen is saturated with content. The human is not served.

The loop is self-perpetuating. Politicians who optimize for engagement get more visibility. Visibility attracts donors and attention. Donors and attention shape policy. The policy that results is designed for the data point, not the human. And the human, saturated and exhausted, disengages. The disengagement is not a failure of the system. It is the system working as designed.

The Exits

The exits are the same as we identified in the broader Data Point Democracy analysis. Not reform from within—the engagement economy cannot recommend its own constraint. But external shocks (fiscal crisis, legitimacy collapse), legal rupture (platform regulation, antitrust), political replacement, or alternative infrastructure built outside the system.

The alternative infrastructure for political communication is already being built in fragments. Independent media. Open-source platforms. Participatory budgeting. Deliberative democracy experiments. Community organizing that operates outside the algorithmic feed.

None of these scale the way the engagement economy scales. That is the point. The engagement economy is fast, cheap, and addictive. The relational alternative is slow, local, and requires presence. It will not out-compete the engagement economy on its own terms. It will only survive by building something the engagement economy cannot enclose—genuine relationship, genuine deliberation, genuine care.

The engagement economy is not eternal. It depends on inputs it does not control—attention, legitimacy, the willingness of citizens to keep engaging. When those inputs fail, the window opens. The work between now and then is preparation. The record. The archive. The alternative.