The Citizen Who Still Has to Think: Artificial Intelligence and the Conditions of Democratic Judgment

Carl Jean

Person at laptop uses AI to generate an argument while a wall of handwritten questions shows the harder process of forming independent judgment.

Abstract

This essay argues that artificial intelligence threatens democracy not primarily through misinformation or manipulation, but by displacing the formative processes through which citizens develop the capacity for independent political judgment. It introduces the concept of civic displacement, defined as the removal or compression of the experiences—exposure to disagreement, evaluation of evidence, and the effort of articulation—through which democratic judgment is formed. While existing accounts emphasize the manipulation of political content and the erosion of shared facts, this essay argues that the deeper threat is structural: the removal of the conditions under which citizens develop the capacity to evaluate content at all. Drawing on the work of John Dewey and Jürgen Habermas, the essay situates artificial intelligence not at the level of political content, but at the level of civic formation itself. When reasoning is simulated rather than enacted, participation can expand even as the processes that sustain democratic capacity are weakened. Both produce participation; only one preserves formation. The result is a form of civic surface formation: the appearance of engagement without the processes through which judgment is developed. The danger is not that citizens will be deceived, but that they will be furnished with political positions they never had to earn.

I. The Citizen and the Interval of Judgment

A citizen sits at a laptop late in the evening, preparing to submit a public comment on a proposed zoning change. The notice has been available for weeks, but the issue is difficult to grasp—technical language, competing claims, uncertain consequences. The deadline is approaching. Rather than working through the material line by line, the citizen opens an artificial intelligence tool and enters a prompt: Summarize the debate and generate a strong argument opposing the proposal.

Within seconds, a response appears. It is coherent, well-structured, and persuasive. It identifies the key issues, anticipates counterarguments, and frames the position in language that is both precise and accessible. The citizen reads it, makes minor adjustments, and submits it under their name.

Nothing in this scene is deceptive. The argument may be accurate. The position may be defensible. The citizen has, in a formal sense, participated in a democratic process. Yet something has changed in the structure of that participation.

What is absent is not information, but an interval.

Between encountering a political question and arriving at a judgment lies a period of uncertainty: the slow effort of trying to understand what is at stake, the difficulty of weighing competing claims, the friction of discovering one's own position by attempting to articulate it. This interval is not incidental to democratic life. It is where judgment is formed. It is where citizens move from exposure to opinion toward the development of a view that is genuinely their own.

In this case, that interval has been compressed—perhaps eliminated altogether. The citizen does not struggle to formulate an argument; the argument is provided. They do not test their assumptions against resistance; the resistance has already been anticipated and resolved. The process of arriving at a judgment is replaced by the selection and endorsement of one.

The result is a form of participation that is procedurally intact and, in many respects, improved. The submitted comment is clearer, more organized, and more rhetorically effective than what the citizen might have produced independently. From the perspective of the institution receiving it, nothing appears to be missing. Yet the process through which political judgment would have been formed has not occurred.

This is not a failure of the citizen, nor is it an instance of manipulation. It is a structural shift in the conditions under which democratic participation takes place. When the work of articulating a position can be delegated to systems that perform it more efficiently, the effort required to develop that position becomes optional. Participation remains; formation becomes contingent.

The form of democratic engagement persists, even as the conditions under which democratic judgment is formed begin to erode.

II. Civic Displacement and the Reorganization of Judgment

The scene described above does not represent a failure of information, nor primarily a distortion of it. The citizen is not misled. The argument they submit may be accurate, even well-informed. What has changed is not the content of political reasoning, but the conditions under which it is produced.

This essay describes that change as civic displacement: the removal or compression of the processes through which citizens develop the capacity for independent political judgment. These processes include exposure to disagreement, the effort of evaluating competing claims, and the difficulty of articulating a position in one's own terms. They are not merely preparatory steps toward participation. They are the conditions under which participation acquires democratic substance.

Existing accounts of artificial intelligence and democracy have focused primarily on the integrity of political content. Concerns about misinformation, algorithmic amplification, and targeted persuasion have emphasized the ways in which citizens may be exposed to false, misleading, or strategically curated information. These analyses are indispensable, but they share a common assumption: that the central problem lies in what citizens encounter.

Civic displacement shifts the focus from content to formation. The question is not only whether citizens are receiving accurate information, but whether they are developing the capacity to evaluate information at all. Artificial intelligence alters this capacity by intervening in the processes through which judgment is formed. When systems can summarize debates, generate arguments, and anticipate objections, they do not merely influence what citizens think—they reduce the necessity of thinking through those questions independently.

This transformation is subtle because it does not prevent engagement. It facilitates it. Citizens can participate more easily, more quickly, and often more effectively than before. The outputs of participation—comments, arguments, positions—may improve in clarity and coherence. Yet the relationship between participation and formation is altered. The work that once connected the two becomes optional.

The consequence is not the disappearance of judgment, but its displacement. Judgment appears in the form of completed positions, but the processes through which those positions would have been developed are externalized. Citizens no longer need to move through the interval of uncertainty described in Section I; they can select from outcomes generated elsewhere.

This distinction becomes clearer when situated within the democratic theory of John Dewey. For Dewey, the public is not a fixed body but something that must be formed through the shared recognition of consequences and the collective effort to address them. Democracy depends not simply on participation, but on the development of citizens capable of inquiry—individuals who can identify problems, evaluate evidence, and revise their views in response to others.

These capacities are not given. They are cultivated through practice. The act of engaging with a political question—struggling to understand it, encountering opposing arguments, attempting to articulate a position—is itself the process through which democratic competence emerges. Remove or compress these processes, and the formation of the public is weakened at its source.

Civic displacement therefore names a structural shift: the decoupling of democratic participation from the processes that historically produced the capacity to participate meaningfully. Participation remains available and may even expand, but the formation it presupposes becomes contingent.

This is not a failure of the citizen, nor is it an instance of manipulation. It is a transformation in the conditions under which political judgment becomes possible.

III. The Public Sphere Under Simulation

If civic displacement alters the formation of individual judgment, its consequences extend further into the structure of democratic life itself. Political judgment is not formed in isolation. It emerges within a shared space of communication in which citizens encounter one another's claims, contest them, and revise their views in response.

This space is what Jürgen Habermas describes as the public sphere: a domain in which private individuals come together to deliberate about matters of common concern. Its legitimacy depends not on consensus, but on the possibility of communicative rationality—the idea that participants engage one another through reasons that can, in principle, be examined, challenged, and defended.

This model presupposes more than the circulation of arguments. It depends on the presence of agents capable of generating, evaluating, and responding to those arguments through their own capacities. The public sphere is not simply a repository of positions; it is a process through which positions are formed, tested, and transformed.

Artificial intelligence introduces a new condition into this process: the possibility that participation in public discourse can occur without the corresponding activity of reasoning. When systems generate arguments, anticipate objections, and produce rhetorically complete positions, the visible structure of deliberation remains intact. Claims are made, counterclaims appear, and positions are articulated with increasing clarity and sophistication.

Yet the underlying activity that gives these exchanges their democratic meaning is altered. The reasons presented in public discourse may no longer be the result of participants working through disagreement, but the output of systems that have already performed that work in advance. What appears as deliberation may instead be the interaction of completed arguments that have not been formed through the processes they represent.

This is not simply a matter of distortion or bad faith. The arguments themselves may be coherent, informed, and responsive to opposing views. The problem lies in the conditions of their production. When reasoning is simulated rather than enacted, the relationship between communication and understanding is weakened.

Habermas's account of communicative action depends on the idea that participants are oriented toward mutual understanding through the exchange of reasons. This orientation presupposes that the act of offering a reason is tied to the capacity to justify it—to explain how it was arrived at, to defend it under challenge, and to revise it when confronted with better arguments.

When arguments are generated externally, this connection becomes uncertain. A citizen may present a position that is, in principle, defensible, without having developed the capacity to defend it independently. The exchange of reasons continues, but the link between reasoning and the reasoner is attenuated.

At the level of the public sphere, this produces a subtle but significant shift. Deliberation persists in form, but its function changes. Instead of serving as a process through which collective judgment is formed, it risks becoming a site in which pre-formed judgments are displayed, circulated, and endorsed.

The effect is cumulative. As more participants rely on systems to generate and refine their contributions, the proportion of discourse that emerges from direct engagement with disagreement decreases. The public sphere becomes increasingly populated by arguments that are structurally complete but experientially unformed.

This transformation does not eliminate disagreement. It may intensify it. Positions can be generated more quickly, articulated more forcefully, and distributed more widely. But the process through which disagreement produces understanding—through the effort of engaging with opposing views—is weakened.

The public sphere continues to operate, but the activity that once gave it democratic substance begins to shift from human agents to the systems that mediate their participation.

Civic displacement therefore operates at two levels simultaneously. At the level of the individual, it weakens the formation of political judgment by compressing the interval in which that judgment would be developed. At the level of the collective, it alters the communicative conditions under which those judgments are brought into relation with one another.

The result is not the disappearance of deliberation, but its simulation: a public sphere in which the exchange of reasons persists, even as the processes through which those reasons are formed increasingly occur elsewhere.

IV. The Counterargument: Access, Inclusion, and the Expansion of Participation

The account of civic displacement developed thus far appears to rest on a demanding conception of democratic participation—one that assumes citizens must engage directly in the difficult work of forming political judgments. Against this, a powerful counterargument can be made: that artificial intelligence does not erode democracy, but expands it.

From this perspective, the primary barrier to democratic participation has never been the absence of information, but the uneven distribution of time, education, and communicative resources. Many citizens lack the capacity to engage deeply with complex policy issues—not because they are unwilling, but because the demands of work, caregiving, and daily life leave little room for sustained deliberation. Others may possess views but lack the linguistic or rhetorical tools to articulate them effectively in formal political settings.

Artificial intelligence, on this view, lowers these barriers. By generating coherent arguments, summarizing complex debates, and translating positions into accessible language, it enables broader participation in democratic processes. Citizens who might otherwise remain silent can now contribute. Those with limited expertise can engage with technical issues. Those without formal training can express their views in ways that are legible within institutional frameworks.

In this sense, AI appears not as a force of displacement, but of inclusion. It democratizes the tools of political expression, making participation more accessible and more equitable. From this perspective, the citizen in Section I is not bypassing democratic engagement, but finally gaining access to it.

This argument has substantial force. It identifies real constraints on democratic participation and offers a plausible mechanism for addressing them. Any account of civic displacement that ignores these benefits risks defending a form of democracy that privileges those who already possess the time, education, and confidence to engage without assistance.

The question, however, is not whether artificial intelligence expands participation. It clearly can. The question is whether the form of participation it enables preserves the conditions under which democratic judgment is formed.

The distinction developed in the previous sections becomes decisive here. Participation can be expanded in ways that either support or displace formation. Tools that assist citizens in understanding issues, encountering diverse perspectives, and refining their own arguments may deepen democratic capacity. Tools that replace the process of forming a position with the selection of a pre-formed one may expand participation while weakening the capacities participation was meant to cultivate.

The difference is not one of degree, but of structure. In one case, assistance operates within the interval of judgment, supporting the processes through which a view is developed. In the other, assistance collapses that interval, presenting the outcome without requiring the process. Both produce participation. Only one preserves formation.

This distinction clarifies why the access argument, though compelling, is insufficient on its own. Expanding participation is not equivalent to strengthening democracy if the expansion occurs through the displacement of the processes that make participation meaningful. A system in which more citizens contribute more frequently may still undermine democratic capacity if those contributions are detached from independently formed judgment.

The expansion of participation does not, by itself, guarantee the preservation of the conditions under which democratic judgment is formed.

The task, then, is not to reject artificial intelligence as a tool for civic engagement, but to distinguish between forms of assistance that cultivate judgment and those that render it unnecessary. Civic displacement names the point at which this distinction is crossed.

V. Case Study: AI-Generated Participation at Scale

The dynamics of civic displacement are not speculative. They are already visible in the increasing use of artificial intelligence to generate public comments, constituent communications, and other forms of political participation at scale. Regulatory agencies and legislative offices have begun to encounter large volumes of submissions that bear the formal characteristics of citizen engagement while raising questions about the processes through which those contributions were produced.

In recent years, agencies such as the U.S. Environmental Protection Agency and the Federal Communications Commission have received millions of public comments on proposed rules. Investigations into these comment periods have documented the presence of large-scale, automated submissions—some generated through coordinated campaigns, others increasingly produced with the assistance of artificial intelligence systems capable of generating unique, well-structured responses that are difficult to distinguish from independently written comments.

From the perspective of institutional procedure, these submissions are often valid. They address the relevant issues, articulate recognizable positions, and conform to the expected format of public input. In some cases, they are more coherent and comprehensive than the comments that citizens would produce without assistance. The system, in other words, receives participation—and receives it in an improved form.

Yet when examined through the framework developed in this essay, a different pattern emerges.

First, process displacement. Citizens no longer need to work through the underlying policy question in order to participate. The effort of interpreting technical documents, identifying relevant concerns, and structuring an argument can be delegated to systems that perform these tasks rapidly and effectively. The interval in which political judgment would have been formed is compressed or bypassed.

Second, deliberative simulation. The comments produced in this way often incorporate counterarguments, acknowledge complexity, and present themselves as the result of careful consideration. They exhibit the formal features of deliberation—engagement with opposing views, structured reasoning, rhetorical balance—without necessarily being the product of a citizen's own engagement with those views. The appearance of deliberation persists, even as its underlying activity is displaced.

Third, judgment substitution. Citizens may endorse and submit positions that they have not independently developed. The act of participation shifts from the formation of a judgment to the selection and authorization of one. Political positions are no longer arrived at through the effort of articulation, but adopted from outputs generated elsewhere.

Taken together, these conditions mark the threshold of civic displacement. Participation remains not only possible, but amplified. Institutions receive more input, more rapidly, and in forms that are easier to process. Yet the relationship between participation and formation is altered. The system captures the output of engagement without necessarily capturing the process that gives that engagement its democratic significance.

The effects are not limited to regulatory comment periods. Similar dynamics are emerging in the domain of legislative representation, where organizations have begun using automated tools to generate large volumes of constituent-style messages directed at elected officials. These communications are often tailored, contextually relevant, and rhetorically persuasive. They resemble the expressions of individual citizens, even when generated through centralized systems.

From the perspective of a legislative office, the result is an influx of apparent public opinion. Messages arrive in large numbers, representing positions on specific issues, often framed in language that suggests personal conviction. Yet the connection between these messages and independently formed citizen judgment is increasingly uncertain. What appears as distributed political will may, in fact, reflect the outputs of systems that have aggregated, generated, and distributed positions at scale.

This is not simply a problem of authenticity or fraud. Even when citizens willingly use these tools, the structure of participation is transformed. The representative receives a signal, but the signal may no longer correspond to the process through which democratic judgment is formed.

Civic displacement becomes visible when systems designed to register public input begin to capture the products of participation without the processes that once produced them.

The significance of this shift lies in its subtlety. Nothing in the formal structure of democratic procedures is necessarily violated. Comments are submitted. Messages are sent. Positions are expressed. Yet the conditions under which those actions acquire meaning are altered. Participation scales. Formation does not.

At this point, civic displacement is no longer a theoretical concern. It is an observable feature of contemporary democratic practice, operating simultaneously at the level of individual engagement and institutional response.

VI. The Civic Threshold and the Conditions of Preservation

The preceding sections have described a structural transformation in the conditions of democratic participation. Artificial intelligence does not eliminate participation, nor does it necessarily degrade the quality of political content. It alters the relationship between participation and the processes through which political judgment is formed.

This transformation converges at a threshold: the point at which democratic participation can occur without requiring the formation of independent political judgment. Below this threshold, technological assistance operates within the interval of judgment, supporting the processes through which citizens develop their views. Beyond it, assistance substitutes for those processes, presenting outcomes without requiring the effort through which they would have been produced.

The civic threshold is defined by the interaction of three conditions already visible in contemporary practice.

First, cognitive outsourcing. Political reasoning is generated externally by systems that can summarize debates, construct arguments, and anticipate objections. Citizens no longer need to engage directly in the work of interpreting issues or developing positions. The effort required to arrive at a view becomes optional.

Second, deliberative simulation. The outputs of these systems exhibit the formal characteristics of deliberation—engagement with opposing views, structured reasoning, and rhetorical balance—without requiring the underlying activity of reasoning itself. What appears as deliberation may instead be the interaction of arguments that have been produced in advance.

Third, judgment substitution. Citizens participate by selecting, endorsing, and transmitting positions generated elsewhere. The act of judgment shifts from formation to authorization. Positions are no longer developed through engagement with disagreement, but adopted from outputs that have already resolved that engagement.

Individually, these conditions do not necessarily undermine democratic capacity. Tools that assist in understanding complex issues or expose citizens to diverse perspectives may strengthen the processes of judgment formation. It is their convergence that is decisive. When cognitive outsourcing, deliberative simulation, and judgment substitution operate together, participation becomes decoupled from formation.

At this point, the distinction developed in Section IV becomes operational. Participation is expanded, but the expansion no longer guarantees the preservation of the conditions under which democratic judgment is formed. Citizens can engage without developing the capacities that engagement was historically meant to cultivate.

The problem is not that artificial intelligence introduces error into democratic processes. It is that it alters the conditions under which those processes produce citizens capable of sustaining them. A democratic system can tolerate misinformation, disagreement, and even strategic manipulation to a degree, because it relies on the capacity of citizens to evaluate, contest, and revise their views. When that capacity is weakened, the system loses the mechanism through which it corrects itself.

The question, then, is not whether artificial intelligence should be used in civic contexts, but under what conditions its use preserves the processes of formation rather than displacing them.

A minimal account of these conditions can be specified.

First, systems must preserve the interval of judgment. They may assist in presenting information or organizing arguments, but they must not eliminate the need for citizens to engage in the work of evaluating and articulating their own views.

Second, systems must sustain exposure to disagreement. Rather than resolving opposition in advance, they should make the presence of competing claims visible and require users to confront them directly.

Third, systems must maintain the link between judgment and responsibility. Citizens should be required to justify the positions they endorse in ways that reflect their own understanding, rather than relying solely on the authority of generated outputs.

These conditions do not prohibit the use of artificial intelligence. They define the boundary between assistance and substitution. Within that boundary, technological tools can support democratic engagement by making complex issues more accessible and by facilitating the exchange of ideas. Beyond it, they risk transforming participation into a process that no longer produces the capacities it presupposes.

Civic displacement, in this sense, is not an absolute state but a gradient. The task is not to eliminate technological mediation, but to prevent its convergence beyond the threshold at which formation becomes unnecessary.

VII. What Must Be Preserved: Practices of Civic Formation

The conditions specified in Section VI define the boundary between assistance and substitution. To remain below the civic threshold, democratic systems must preserve the interval of judgment, sustain exposure to disagreement, and maintain the link between judgment and responsibility. These conditions are not merely theoretical. They imply concrete practices in the design of civic technologies and the structure of civic education.

Consider first the design of civic participation tools. Many existing systems prioritize efficiency: summarizing issues, generating arguments, and enabling rapid submission of positions. A system designed to preserve formation would operate differently. Rather than presenting a completed argument in response to a prompt, it would structure the interval of judgment itself.

Such a system might require users to engage with multiple competing claims before articulating a position. It could present arguments in tension, without resolving them, and prompt the user to identify points of agreement and disagreement. Instead of generating a finished comment, it could guide the user through stages of articulation—first identifying concerns, then formulating reasons, and only then composing a final statement. The system would assist in organizing thought without substituting for it.

Similarly, the submission process itself could be structured to preserve responsibility. Rather than allowing users to endorse generated outputs directly, systems could require justification in the user's own terms: a brief explanation of why a position is held, or a reflection on how competing arguments were evaluated. These requirements would not guarantee the depth of engagement, but they would reintroduce the necessity of articulation that civic displacement removes.

The same principles apply at the level of civic education. If democratic capacity depends on the formation of judgment, then education must be oriented not only toward informing students, but toward requiring them to think. This requires a shift from teaching political content to designing conditions in which the processes of judgment cannot be bypassed.

In practice, this means structuring civic learning around tasks that demand engagement with disagreement. Students must encounter opposing arguments not as summaries, but as positions that resist easy resolution. They must be required to articulate and defend their views in contexts where those views can be challenged. Artificial intelligence can assist in this process by expanding access to information and perspectives, but it must not remove the necessity of working through them.

This approach parallels the principle developed in the essay on instructional design: thinking must be made structurally necessary. At the civic level, this means designing institutions and technologies such that participation requires the exercise of judgment rather than the selection of pre-formed positions.

These practices do not eliminate the use of artificial intelligence in democratic contexts. They redefine its role. Instead of producing political reasoning, systems would support the conditions under which reasoning occurs. Instead of collapsing the interval of judgment, they would make it more accessible and more visible.

The aim is not to prevent citizens from using intelligent systems, but to ensure that those systems do not render the work of judgment unnecessary.

Civic displacement can be mitigated only to the extent that the processes it displaces are deliberately preserved. Without such preservation, the expansion of participation risks becoming a transformation in which democratic form is retained while democratic capacity is gradually diminished.

Conclusion

A citizen sits at a laptop, preparing to submit a public comment. The issue is complex. The deadline is near. The tools available can generate a position in seconds—clear, persuasive, and complete.

Nothing prevents the citizen from using those tools. Nothing guarantees that doing so will undermine democratic life. The argument may be sound. The position may be justified. The process may even appear, from the outside, to exemplify effective participation.

What the preceding analysis has shown is that something more fundamental is at stake. The question is not whether citizens can produce political positions, but whether they must still develop them. Between encountering a question and arriving at a judgment lies the interval in which democratic capacity is formed. When that interval is compressed or bypassed, participation persists, but the conditions that give it meaning begin to erode.

This is the logic of civic displacement. It does not remove the structures of democracy. It reorganizes the conditions under which those structures produce citizens capable of sustaining them. At the level of the individual, it weakens the formation of judgment. At the level of the public sphere, it transforms deliberation into the circulation of pre-formed positions. At the level of institutions, it captures the outputs of participation without securing the processes that once produced them.

The danger is not that citizens will be deceived. It is that they will be furnished with political positions they never had to earn.

Democratic systems have always depended on an assumption that is rarely made explicit: that citizens are capable of thinking through the questions that confront them. This capacity is not guaranteed. It is produced through practices that require effort, expose individuals to disagreement, and demand the articulation of reasons.

Artificial intelligence does not eliminate this capacity. It renders it optional.

Participation scales. Formation does not.

A democratic system remains viable only to the extent that it preserves the conditions under which citizens must still think.

Join the Conversation

What happens when the systems we rely on begin to shape not only what we produce, but how we arrive at our judgments? If the interval between encountering a question and forming a view is compressed or bypassed, what becomes of the capacities that democratic participation presupposes?

If you found this argument compelling—or if you disagree—I invite you to share your perspective in the comments. How is artificial intelligence affecting the way political judgments are formed in your experience? Subscribe to follow the full series as it continues to examine the changing conditions of expertise, responsibility, and civic life in the age of intelligent systems.

Works Cited

Crawford, Kate. Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence. Yale University Press, 2021.

Dewey, John. The Public and Its Problems. Henry Holt, 1927.

Habermas, Jürgen. The Structural Transformation of the Public Sphere. Translated by Thomas Burger, MIT Press, 1989.

New York State Office of the Attorney General. The Fake Comments Crisis. 2021.

Pariser, Eli. The Filter Bubble: What the Internet Is Hiding from You. Penguin Press, 2011.

Pasquinelli, Matteo. The Eye of the Master: A Social History of Artificial Intelligence. Verso, 2023.

Starbird, Kate, et al. “Ecosystem or Echo-System? Exploring Content Sharing across Alternative Media Domains.” Proceedings of the International AAAI Conference on Web and Social Media, vol. 12, 2018.

Tufekci, Zeynep. Twitter and Tear Gas: The Power and Fragility of Networked Protest. Yale University Press, 2017.

Search This Blog

The Carl Jean Journal