Platform Power

The Philosophical Object of Platform Power

Platform power arises when an actor that organizes the conditions of interaction thereby acquires the capacity to shape the conduct of the participants. Digital platforms connect users, sellers, developers, advertisers, workers, institutions, or information sources through rules and technical interfaces. They often create genuine opportunities for coordination. Their political significance emerges because participation can depend upon terms that participants have limited ability to negotiate, evaluate, or change. The philosophical question concerns the institutional position of the intermediary: who establishes the conditions of access, which interests count in their formulation, what recourse exists when the terms change, and who can legitimately challenge the resulting authority?

An online marketplace can decide which sellers appear prominently, which payment methods are admissible, what constitutes prohibited conduct, whether access may be suspended, and how records of performance are interpreted. A social communication service can classify posts, recommend material, restrict discoverability, and decide whose appeal deserves attention. A mobile operating system or application store can govern technical compatibility and impose distribution terms. These activities have different functions, but all can place the power to organize possibilities in a relatively small number of institutional hands. The individual transaction is often intelligible only against the background of this ongoing power to construct the transaction's conditions.

This study uses platform as an analytical category for a socio-technical intermediary that enables interactions among distinct participants while retaining material control over at least some rules, interfaces, data flows, or conditions of access. Platforms differ in ownership, sector, technological organization, legal status, and the extent of their infrastructural dependence. A university learning system, an open-source federation, an employment-matching service, and a dominant advertising marketplace are not interchangeable political institutions. The concept identifies a family resemblance; the analysis of power asks what particular capacities follow in particular arrangements. A platform need not be dominant for its moderation rules to influence speech, nor must every highly influential platform exercise arbitrary domination.

A preliminary definition follows. Platform power is the relational capacity to establish, modify, or enforce consequential conditions of mediated participation, including the structure of admissible actions, their visibility, their rewards, and the practical costs of refusing those conditions. The definition concerns capacities and relationships. It does not assume that every use of the capacity is coercive, that every outcome is intended, or that every technically produced difference is unjust. It makes it possible to ask whether participants can understand, contest, escape, or collectively alter the conditions affecting them.

The boundaries of this dossier follow from the definition. General political power, the nature of algorithms, competition law, digital privacy, the theory of sovereignty, and the ontology of artificial agents each deserve independent investigation. Here they enter only when they explain the political structure of platforms. The distinctive explanandum is control through the design and administration of shared digital environments. This control can become publicly consequential even when exercised through a contract between private parties.

Power, Authority, Influence, and Dependence

The word power is philosophically ambiguous because it may refer to the possession of an ability, the successful production of an outcome, a recognized authority, or a social relation in which one party's possible actions depend upon another. These senses overlap, but they support different judgments. A platform may have the technical capacity to remove a profile without using it; it may influence behavior through recommendations; it may claim an authority to enforce community standards; and it may possess economic leverage because participants cannot leave without serious loss. A satisfactory assessment must distinguish the capacity, its exercise, its justification, and its consequences.

Thomas Hobbes's Leviathan (1651), especially chapters 17–18 and 21, provides an early modern account of how rules, authorization, and the securing of collective order are related. His theory addresses a political commonwealth constituted through authorization rather than a private digital service. The relevant comparison is structural: an authority can settle disputes and coordinate action because participants recognize an organizing center, yet the legitimacy of that organizing center cannot be derived simply from its ability to impose rules. In Hobbes, the conditions of authorization and political obligation are central to the account; using his vocabulary for platforms therefore requires an explicit account of the differences between political founding, membership, and ordinary terms of service.

Max Weber's analysis of legitimate domination in Economy and Society (first published posthumously in 1922) clarifies why organized administration has a specifically social basis. Stable rule involves expectations about commands, compliance, offices, and procedures. A platform's automated enforcement and bureaucratic moderation can resemble rationalized administration in their calculability and reliance on impersonal procedures. Nevertheless, algorithmic regularity by itself establishes neither the legitimacy of the norms enforced nor the right of an organization to administer them. A consistent but unfair procedure may reproduce arbitrary exclusion more efficiently.

Hannah Arendt's On Violence (1970) distinguishes power grounded in people acting together from violence understood in instrumental terms. This illuminates a feature that purely corporate accounts obscure: participants can form publics, communities, and movements that give a platform much of its value. The capacity of the company to dissolve such relations is different from the collective power generated by the persons using its facilities. Indeed, a platform can profit from relationships it did not create while retaining the final authority over their technical accessibility. The tension is particularly acute when collective associations become dependent on one commercial infrastructure.

Modern political philosophy adds two necessary distinctions. Influence can alter opinions without restricting admissible actions. Domination, in Philip Pettit's republican account, concerns a form of subjection to uncontrolled or arbitrary power even where intervention has not occurred (Pettit, 1997). A user who has never been suspended may nevertheless be dependent upon an operator able to change rules or revoke access without adequate constraint. Conversely, an operator that imposes strict moderation subject to intelligible rules, fair procedures, and independent review may interfere with a user while remaining less dominating than a formally permissive but entirely discretionary one. This distinction helps avoid equating all restrictions with losses of freedom.

Power therefore admits at least four assessments in the platform setting: the scope of an operator's capacity; how and when the capacity is actually exercised; the degree of participant dependence; and the normative status of the operator's claim to rule. These questions should be investigated together because the same intervention can have different moral significance depending upon who is affected, what alternatives exist, and what mechanisms of redress are available.

From Public Communication to Technically Mediated Publics

The modern public sphere provides a genealogy of the political problem. Jürgen Habermas's The Structural Transformation of the Public Sphere (1962; English translation by Thomas Burger, 1989) reconstructs the emergence and transformation of institutions in which persons discuss matters of common concern. Its central relevance is the institutional mediation of public reasoning: access to information, the organization of publicity, social exclusion, and the pressures of commercial structures affect what appears as public deliberation. The platform question emerges when access to such mediated publicity increasingly depends upon privately owned and continuously modified systems.

Nancy Fraser's 1990 critique of the actually existing bourgeois public sphere challenges the idea that a single officially inclusive arena necessarily provides equal participation. Unequal standing, the construction of purportedly private interests, and the formation of counterpublics complicate the historical ideal. A digital platform may host a great many speakers while still reinforcing inequalities of discoverability, safety, influence, or bargaining position. The democratic question is consequently wider than whether an account may technically post a message. One must ask whether marginalized speakers can sustain counterpublics, acquire reliable attention, and resist rules shaped by already dominant participants.

Traditional media organizations controlled particular channels of circulation; digital platforms often also control the technical organization of the interactions among audiences, producers, advertisers, and other media institutions. Networked communication can expand participatory capacity while creating centralized points of control over visibility and distribution. The result is an institutional hybrid: private services provide settings for activities with clear public significance while their technical mechanisms remain within corporate infrastructures.

Jack Balkin's account of freedom of expression as a triangle (2018) is useful here. Free speech disputes involve not only speakers and territorial states but infrastructure providers whose decisions shape the ability to communicate. Governmental demands, corporate content policies, advertisers, users, and civil organizations interact rather than standing in a simple public/private opposition. Daphne Keller's examination of hybrid state-platform power (2019) shows how government pressure and intermediary discretion can combine in ways that make legal responsibility harder to identify. These analyses caution against two simplifications: treating a private platform as a sovereign state, and treating the platform as a passive venue unrelated to public speech.

The normative object is an environment of public communication whose rules arise through overlapping institutions. Democratic access must therefore be assessed at the level of practical possibilities: who can speak and be heard, who sets the rules of intelligibility and reach, and whether exclusions can be justified to those bearing their effects.

Artifacts, Architecture, and the Political Significance of Design

Langdon Winner's "Do Artifacts Have Politics?" (1980) develops the claim that technical artifacts and systems may embody forms of power and authority. The argument includes two distinguishable possibilities: designs that settle particular social disputes and technological arrangements whose operation is compatible with, or tends to require, particular political relations. Digital platforms make the question especially direct. Default settings, permission hierarchies, interoperability standards, and allocation of access rights affect what participants can do before any moderator makes an explicit decision.

An interface distributes effort and attention. A ranking system constitutes a practical ordering of the visible. An application programming interface establishes which actors may connect, what kinds of information may pass, and who can withdraw compatibility. A payment or identity system fixes conditions for entry. These arrangements can implement norms that are expressed in prose, but they can also incorporate choices that are not presented as explicit normative decisions. Their political significance becomes clear when a formally available action is practically hidden, unreasonably difficult, or dependent on opaque permission.

Lawrence Lessig's Code: Version 2.0 (2006) offers a sustained account of the regulatory effect of software architecture. Law, market pressures, social norms, and code interact as modes of constraint. The philosophical contribution is the recognition that constraints are partly enacted in the structure of available action itself. Rules can be implemented as enforcement after a violation; they can also be built into the interface so that the prohibited action cannot be performed within the system. The latter arrangement may reduce discretionary enforcement while rendering the rule less salient and harder to contest.

The observation does not imply technological determinism. Design is produced within commercial, legal, cultural, and professional circumstances. A recommender system's technical operation is inseparable from decisions concerning what counts as engagement, whose experience is optimized, and which organizational objectives are treated as relevant. The same generic technology can serve public-interest, commercial, and coercive purposes depending upon institutional arrangements. This is why Winner's problem belongs to political philosophy as well as engineering: a design choice can establish a durable distribution of possibilities without acquiring democratic legitimacy through its technical implementation.

The political evaluation of architecture requires more than exposing source code. Important questions concern objective functions, acceptable errors, availability of meaningful alternatives, governance over future changes, and the possibility of collective refusal. An open algorithm can still govern unfairly where its training data, implementation incentives, or institutional setting remain unaccountable. Conversely, a partly proprietary system may be subject to robust external testing and rights-protective oversight. Transparency is a means of understanding and contesting power, not a sufficient theory of legitimate power.

From Discipline to Modulation: Foucault and Deleuze

Michel Foucault's analysis of power develops the idea that social conduct is shaped through relations, practices, forms of knowledge, and the organization of possible action. In "The Subject and Power" (1982), power is examined through actions upon the actions of others rather than restricted to the ownership of a resource or the transmission of an order. This approach is particularly relevant to platforms because they can structure the field of possible conduct through rankings, norms, interface design, incentives, sanctions, and persistent evaluation.

A platform operator need not issue an individualized command to each participant. Rules can be expressed in a general policy, translated into decision procedures, and applied through automated or human review. The participant anticipates these mechanisms and adapts behavior. Creators modify language to avoid sanction; sellers alter practices in response to metrics; workers manage ratings; developers comply with interface permissions. These anticipatory adjustments can be rational responses to an environment even if their individual causes remain difficult to observe. The same architecture creates opportunities and normalizes certain forms of self-management.

Gilles Deleuze's "Postscript on the Societies of Control" appeared in French in 1990 and in English translation in October in 1992. The essay describes a historical transition from enclosed disciplinary institutions toward flexible, continuous modulation. Its concepts offer an interpretive vocabulary for platform environments in which access, ranking, authentication, and reward change continuously. The platform subject can be assessed through profiles, risk scores, recommendation parameters, or changing qualifications rather than through one static rule. This analogy is philosophically suggestive; the short essay did not predict particular present-day platform business models, and its historical thesis should not be treated as an empirical demonstration concerning every digital service.

Foucault's approach illuminates the productive character of power: platforms do not merely restrict an antecedently complete field of social activity; they create identities, audiences, occupations, categories, and forms of visibility. Deleuze highlights the temporal dimension: authority can operate through continuous adjustment, thresholds, permissions, and recoding. Together they make a crucial conceptual point. The political question is not exhausted by identifying who commands whom. It includes how a continuously administered system constructs the conditions under which participants come to perceive, evaluate, and exercise their capacities.

The Rise of Platform Studies and the Transformation of Intermediation

Tarleton Gillespie's "The Politics of 'Platforms'" (2010) examines how the word platform helps internet companies present themselves to different audiences. The term connotes technical infrastructure, an opportunity to speak, and a neutral supportive base, even while platform operators make consequential editorial and economic decisions. This linguistic analysis supplies an indispensable methodological warning: self-description is not a theory of institutional power. A service's claim to be a neutral intermediary should be tested against its actual rules, incentives, and practices.

The history of online intermediaries cannot be explained by technology alone. Nick Srnicek's Platform Capitalism (2017) places the expansion of platforms within the political economy of capitalism, data capture, competitive concentration, and changing business strategies. Platforms can aggregate interaction and make their position more valuable as participation grows. Network effects, scale economies, switching costs, and control of complementary services can transform an initially contestable intermediary into a consequential gatekeeper. These possibilities are sector-specific and cannot be reduced to a universal law that all platforms become monopolies.

José van Dijck, Thomas Poell, and Martijn de Waal's The Platform Society (2018) examines how public values become entangled with mechanisms of datafication, commodification, and selection across sectors. The framework broadens the research object beyond social media: education, transportation, health, and news can become dependent upon commercial intermediation. The philosophical problem then concerns public values within hybrid infrastructures. If a school selects a platform for instruction, learners may be subject to practical and informational conditions that they had little role in authorizing. If public services depend on private identification or hosting systems, changing vendor rules can affect the effective accessibility of public institutions.

Jean-Christophe Plantin and colleagues argue that infrastructure studies and platform studies illuminate convergent tendencies: platforms acquire infrastructural characteristics while infrastructures can adopt platform logics (Plantin et al., 2018). Susan Leigh Star's work on infrastructure (1999) had already shown that infrastructures are relational, situated, and often most visible at points of breakdown. A platform becomes infrastructurally significant when its ordinary availability is presupposed by other social practices. The loss of access then interrupts relationships and activities beyond the immediate service.

The conceptual transformation is from intermediation as a useful service to intermediation as a condition of participation. This does not occur uniformly or irreversibly. Open protocols, multi-homing, institutional procurement, competition, collective organization, and public alternatives can preserve countervailing capacities. But where effective alternatives diminish, the power to set platform conditions becomes a form of authority over practices that extend beyond the owner's internal affairs.

Political Economy, Information, and Legal Construction

Julie E. Cohen's Between Truth and Power (2019) treats informational capitalism as a historically developing legal and institutional order. Property, contracting, informational extraction, and governance are shaped by law rather than simply appearing outside it. This argument matters because platform control is frequently presented as the natural consequence of private ownership. In fact, ownership, intellectual property, privacy law, contract enforcement, competition rules, labor classification, and state procurement help establish the powers available to an intermediary. The legal environment participates in constructing the market power it later attempts to regulate.

Shoshana Zuboff's The Age of Surveillance Capitalism (2019) analyzes business practices organized around behavioral data extraction, prediction, and efforts to influence behavior. Her argument helps identify a distinctive source of informational asymmetry. When an operator can observe interactions across a large mediated environment and infer patterns unavailable to the participants, it acquires capacities for prediction, targeting, and intervention. These capacities can be economically valuable and politically consequential. The stronger claim that one particular business logic explains all platform systems must be evaluated case by case: subscription, cooperative, public, advertising-funded, and infrastructural services differ considerably.

Power also appears in the distribution of labor and risk. Alex Rosenblat and Luke Stark's 2016 empirical study of ride-hailing drivers shows how asymmetries of information and platform-mediated management can produce indirect control over ostensibly independent participants. This case is valuable because it demonstrates that formal choice among tasks does not establish practical independence where a company controls access, performance evaluation, compensation signals, and continued participation. The relevant philosophical distinction is between freedom inside an option structure and power over the construction of the option structure.

The fundamental conclusion of this historical reconstruction is that platform power has no single origin in one invention or one author. It gathers older problems of political authority, economic dependence, technology, public communication, administrative rationalization, and control of information into a new institutional configuration. The problem becomes distinct when the organizer of a shared environment can continuously revise the conditions through which other actors obtain access to socially valuable possibilities.

Rule-Making as an Institutional Capacity

A platform's most visible governing activity is the promulgation and enforcement of rules. Terms of service, community standards, seller policies, developer agreements, service-level commitments, and identity requirements constitute an institutional order. The operator defines which participants may enter, what they may do, what evidence establishes violation, which sanctions follow, and how disputes are resolved. Such regulations can be necessary for an environment to function. They also allocate rights and burdens among parties with unequal access to the rule-making process.

Kate Klonick's "The New Governors" (2018) reconstructs the processes through which major social media companies develop and administer speech rules. Its importance for political philosophy lies in the institutional analogy, which must be handled carefully. Moderation resembles governance because it employs rules, trained decision makers, appeals, and regularized enforcement; the platform's rule-making authority nevertheless differs in origin and legal scope from a state's. A contract may generate obligations among parties without amounting to democratic authorization of a public authority. Conversely, democratic states can impose requirements on private intermediaries without automatically becoming rightful arbiters of every editorial decision.

Rule-making power has a temporal dimension. A service can revise the terms governing practices that have already become central to a participant's life or livelihood. A merchant may accumulate customers and ratings on a marketplace before the operator changes commissions or search ranking. An association can build an audience under one moderation regime and confront new restrictions after its members have become dependent on that audience. Consent to the initial terms does not resolve the normative status of all later changes, particularly when affected actors cannot negotiate, predict, or effectively avoid them.

A central criterion is the availability of an intelligible justification. Rules can legitimately protect safety, prevent fraud, preserve interoperability, secure privacy, and establish conditions of productive cooperation. Yet the ends used to justify a rule should be identified and assessed alongside its distribution of burdens. A blanket access restriction can protect a community or exclude a marginalized group; a verification requirement can reduce impersonation or impose insurmountable costs on vulnerable users. The philosophical question concerns whether the operator can provide reasons proportionate to the rule's effects and revise the rule in response to evidence of unjustifiable harm.

The law-like appearance of platform governance should therefore be evaluated through institutional functions rather than rhetorical similarity. Rules become politically important when their scope extends into socially necessary activities, when their effects are extensive, and when those affected have little effective role in their formation or contestation. The capacity to write the rules of access differs from merely participating within them. It is a power over others' powers.

Visibility, Ranking, and the Organization of Attention

A communication platform does not merely decide which messages can remain online. It can determine which messages are encountered, by whom, in what order, and with what cues of prominence or credibility. The ranking of information is a form of practical selection. A user can be formally permitted to speak yet rendered largely inaudible through changes in recommendation or search. Conversely, an operator can amplify material without publicly endorsing it. The capacity to allocate visibility gives the platform influence over opportunities for recognition, conversation, economic exchange, and public judgment.

There is no neutral ranking that displays all available information with equal attention. Scarcity of attention requires selection, and selection presupposes criteria. Chronological order privileges recency; personalized ranking privileges predicted relevance; commercial placement privileges advertising or conversion; editorial ranking privileges norms of importance or quality. Each scheme can be defensible for particular purposes. The political question concerns which purposes deserve priority, how they interact, and what rights participants have when ranking affects essential opportunities.

Safiya Umoja Noble's Algorithms of Oppression (2018) shows how search results and commercial information practices can reproduce socially consequential racialized and gendered representations. The lesson is that ranked outputs must be assessed in their social and historical setting; abstract accuracy cannot by itself settle whether a system undermines equal standing. Different forms of discrimination may operate through training data, business objectives, categorization, and unexamined assumptions about the meaning of relevance. Critical evaluation requires methods for identifying who is disadvantaged, how patterns are produced, and whether remedial changes preserve legitimate communication.

Ranking affects knowledge as well as opportunity. Repeated exposure may form a person's informational environment, while the source and logic of selection remain less visible than the selected content. An epistemic injustice can arise when the design systematically reduces the discoverability or credibility of certain forms of testimony. Yet the reverse danger is equally important: a rule requiring equal display of every claim would erase legitimate distinctions among reliable information, coordinated harassment, and misleading content. The normative objective is an accountable pluralism of information organization, with special attention to consequences for participants whose practical interests depend on visibility.

Opacity complicates the assessment because exposure depends on interactions among data, model behavior, user histories, and organizational experimentation. A public explanation of general ranking principles is valuable, but meaningful accountability also requires evaluation of real distributions of exposure, access to suitable audit evidence, and institutional capacity to correct systemic biases. A platform may satisfy formal disclosure requirements while retaining unilateral authority over the most consequential variables.

Content Moderation and the Conditions of Public Speech

Content moderation includes decisions to remove, label, demote, restrict, or retain material and accounts. The field is normatively complex because platform operators must respond to incompatible claims. People seek protection against harassment, exploitation, fraud, and threats; they also demand opportunities to dissent, criticize, report, organize, and challenge power. No practicable large-scale system can avoid error. The political problem is how to administer unavoidable judgments under rules that preserve freedom, safety, equality, and accountable process.

Tarleton Gillespie's Custodians of the Internet (2018) examines the hidden and often labor-intensive practices of moderation. The study undermines an image of platforms as passive hosts. It also shows why moderation decisions cannot be treated as purely technical classification. Context, meaning, humor, quotation, local conditions, newsworthiness, and asymmetries of vulnerability affect the interpretation of content. Automated models may assist with scale while misreading these relations; human review provides contextual judgment but introduces other inconsistencies and constraints.

The familiar model of individualized due process contains a genuine normative insight: users should normally know what rule was applied, receive an explanation proportionate to the harm, and have effective avenues of correction. Evelyn Douek's "Content Moderation as Systems Thinking" (2022) demonstrates a limit of approaching moderation as a vast sequence of miniature judicial hearings. At enormous scale, individual appeals alone cannot diagnose the institutional causes of error. Organizations should be evaluated through ex ante risk assessment, reporting, system-level evaluation, and oversight of the incentives that produce recurring mistakes.

These approaches are complementary when their purposes are distinguished. Individual review protects a person against the immediate consequences of an erroneous decision. Systemic review asks whether categories, data, incentives, and policies produce patterned harms across populations. A reliable institution needs both. Removing a wrongful sanction without correcting the mechanism that produces thousands of analogous sanctions provides incomplete accountability; changing a system without giving an affected person recourse can similarly leave injustice unaddressed.

State involvement introduces further issues. Authorities may legitimately demand action against illegal content under valid law, yet informal pressure can induce private censorship without the procedural safeguards expected of governmental restriction. Keller (2019) identifies how hybrid state-platform action can make attribution and legal remedy difficult. Platform independence from arbitrary state control is itself an important freedom; platform independence from all public scrutiny would be an equally implausible principle where significant public interests are at stake. The task is to construct conditions under which neither state nor corporation can bypass accountability by relying upon the other.

Data Collection, Classification, and Asymmetric Knowledge

The capacity to structure interaction gives platform operators access to evidence about the behavior of participants. This access can improve coordination, security, recommendations, and service performance. It can also create an informational hierarchy in which an operator knows patterns of the participants' activities that the participants cannot see in equivalent form. Data extraction becomes a political issue when it enables the setting of consequential terms, targeted persuasion, price differentiation, behavioral experimentation, or differential access without meaningful opportunity for affected persons to understand or challenge the classification.

Zuboff's analysis of surveillance capitalism (2019) emphasizes the conversion of behavioral information into predictive and commercially actionable products. The normative significance of this account does not depend on a claim that all data processing is surveillance. Data can be necessary for accessibility, fraud detection, and effective service administration. The relevant distinctions concern purpose, proportionality, retention, combining of contexts, participant agency, and institutional use. Contextual privacy theory offers one way to articulate them: Helen Nissenbaum's Privacy in Context (2010) examines the appropriateness of information flows relative to the social norms of particular contexts.

The informational imbalance can intensify other forms of dependence. A platform may know which alternatives a merchant could profitably use, which customers are likely to leave, or how workers respond to particular incentives. Participants may know only the immediate interface and the apparent consequences of their decisions. Even absent deliberate manipulation, information asymmetry can weaken bargaining and obscure the effects of the company's choices. A policy that allows participants to download personal records may address part of the problem while leaving undisclosed the relational models and comparative assessments that guide the operator's conduct.

The critical task is to identify how knowledge becomes control. Merely possessing information does not necessarily confer political power. The power arises when data-informed classifications alter the feasible opportunities, exposure, sanctions, prices, or social standing of others. The test is institutional: what decisions are made using the information, who may question those decisions, what independent evidence exists, and how participants can protect themselves against misuse.

Economic Gatekeeping, Network Effects, and Exit

Platforms often mediate markets with multiple participant groups. A marketplace brings sellers and buyers together, a developer ecosystem connects application creators to device users, and an advertising exchange connects producers of attention with advertisers. Their value can rise as participation grows, though network effects vary substantially across services and markets. When participants invest in reputations, audiences, compatible products, or specialized workflows, exit may carry costs that strengthen the intermediary's bargaining position.

Nick Srnicek's account of platform capitalism situates this accumulation of intermediary power in business strategies shaped by competition, data accumulation, and the struggle to occupy valuable positions between other actors. The platform can control ranking, commissions, application access, contractual terms, or the transferability of records. The ability to set these rules is distinct from a platform's mere share of transactions. A small specialized platform can be vital for a narrow community, while a large provider may face substantial competitive discipline in some services. Economic power must therefore be assessed by relevant dependence rather than size alone.

To examine exit, distinguish nominal availability from effective feasibility. One competitor may exist in theory, but an independent creator cannot automatically transfer an audience, a trader cannot instantly replicate a history of ratings, and a worker may lose access to information crucial for securing alternative employment. Account portability, interoperability, standardized records, and multi-homing can reduce some of these costs. Their design can also introduce risks of privacy breaches, fraud, or security vulnerabilities. Institutional evaluation needs to compare benefits with these costs rather than treating interoperability as an unconditional solution.

A useful counterfactual asks what would happen if the operator changed an important term unilaterally. Could affected participants obtain a similar service without disproportionate loss? Could they bargain collectively? Could they retain established associations? Could public institutions provide alternative channels? The more serious the consequences of refusal, the stronger the operator's effective authority. This is why market competition and political freedom intersect: inadequate alternatives can convert a formally voluntary exchange into a relationship of structural subordination.

Labor, Algorithmic Management, and Ratings

Platform labor displays the problem in especially concrete form. Digital systems may allocate tasks, direct work, measure performance, manage customer interactions, determine eligibility for assignments, or suspend access. The technical form of management can obscure familiar questions about control, remuneration, risk, employment status, and worker representation. Rosenblat and Stark (2016) investigate the experience of drivers whose formal independence coexists with substantial information asymmetry and indirect managerial control.

The question is how to understand autonomy under algorithmic management. A driver may choose when to log in while remaining dependent on an operator's pricing structure and access rules. A creator may choose what to publish while adapting content to ranking signals and monetization thresholds. A seller may own inventory yet depend on metrics and verification rules that determine whether inventory is effectively visible. Autonomy is therefore multidimensional: discretion over one decision does not erase dependence in other dimensions.

Ratings are particularly revealing. They convert distributed evaluations into numerical signals that can guide future allocation and discipline. Their informational value is real: participants may prefer reliable evidence of performance. But evaluation systems also create opportunities for strategic retaliation, discrimination, context loss, and asymmetries in the ability to challenge judgments. An institution using ratings to impose sanctions acquires responsibilities concerning data quality, appeal procedures, appropriate thresholds, and protection against foreseeable misuse.

Collective organization can alter these relations. Workers, developers, sellers, and creators may coordinate to contest unilateral changes. Yet coordination itself can depend on the same infrastructure whose policies are being challenged. A platform's control over internal messaging, account access, and discoverability may shape the practical possibility of opposition. The deeper political issue concerns whether people whose activities are organized by a common system can form effective countervailing institutions within or outside that system.

Infrastructural Authority and the Layered Digital Order

Not all platforms are primarily visible consumer-facing services. Cloud hosting, identity services, software repositories, payment systems, mobile operating systems, app stores, and distribution networks can acquire influence because other institutions build their operations around them. Their authority is often indirect: a change in a technical standard, application rule, authentication requirement, or contractual condition propagates through dependent services. The relevant political community may include people who have no direct contractual relationship with the infrastructural operator.

Plantin and colleagues (2018) emphasize the convergence of platform and infrastructure characteristics. Star's relational account of infrastructure (1999) explains why dependence may remain invisible until a service fails or a contested rule exposes its operation. Benjamin Bratton's The Stack (2016) offers a wider political and design framework for planetary-scale computation, indicating how layers of computation and geopolitical authority can overlap. The important lesson is that platform power may be distributed across a stack of systems rather than held by one centralized intermediary.

Layering complicates accountability. A forum may moderate content, its hosting provider may impose service policies, a registrar may govern its domain, payment providers may restrict its financial operations, and an app store may determine its availability on devices. Each actor can claim a limited technical role, while their combined decisions determine whether an association can function. A participant may be able to appeal a local moderation decision while lacking recourse against a critical upstream restriction. Political analysis must follow the chain of practical dependence.

The appropriate unit of evaluation is often an institutional ecosystem. Effective contestation requires the ability to identify where a restriction originated, which organization can reverse it, what grounds justify it, and whether comparable alternatives are feasible. An accountability structure that focuses only on the most visible platform will miss upstream power over its possibilities. Conversely, treating every component in the technical stack as identical to a public authority would erase the differences among hosting, security, speech moderation, payment processing, and market intermediation.

The Conversion of Platform Capacity into Political Power

The earlier mechanisms are analytically distinct, but their interaction explains the strength of platform power. Consider a platform that controls admission, ranking, payment, records of reputation, and dispute resolution. Each capacity supports the others: ranking affects earnings; earnings increase dependence; dependence reduces the feasibility of exit; retained data inform further rule changes; and the rule changes become enforceable through control of access. This mutually reinforcing relationship can create a governing position stronger than the sum of individual technical functions.

A formal argument can now be reconstructed. First, political power includes capacities that shape the conditions under which people can pursue, defend, or contest important common interests. Second, platforms can control such conditions through admission, rules, visibility, information, and the organization of economic participation. Third, in circumstances of substantial social dependence, those controls determine opportunities extending beyond a private owner's internal affairs. Therefore, the power exercised by some platforms can be political in effect and normative significance, even when the organizations remain private and their particular conduct is regulated through ordinary corporate and contractual law.

The conclusion is conditional. It does not claim that every platform performs a governmental function or that platforms are states. It identifies the features that trigger political evaluation: consequences for basic or publicly important opportunities; control concentrated in an intermediary; structural dependence; and weak mechanisms through which affected participants can shape or challenge the terms of control. When these conditions are present, a theory of legitimate private exchange alone cannot settle the question.

The next philosophical task is to determine what can justify such control. Ownership, consent, efficiency, safety, and expressive freedom all provide reasons in particular contexts. Each must be evaluated in relation to the different forms of power at issue and to the interests of everyone who is materially affected.

Legitimacy: The Difference Between Having Power and Being Entitled to Exercise It

A platform's capacity to govern participation and its entitlement to govern participation are different questions. A company may lawfully own servers, hold intellectual property, and make agreements with participants without thereby obtaining unlimited moral authority over the social relations that depend upon its infrastructure. Property rights protect important liberties and investments; they operate within a legal and political order that defines both their scope and their limits. The philosophical question is what additional justification becomes necessary when private control affects interests traditionally associated with political freedom, equality, livelihood, association, and public participation.

Legitimacy can be understood as the justified standing of an institution to issue and enforce rules within a specified domain. The domain matters. An organization that offers a specialized creative service may reasonably establish rules protecting its purpose. A platform mediating essential public communication may face stronger demands for fairness and contestability. A payment intermediary may properly impose robust anti-fraud controls while needing protections against arbitrary exclusion. The requirements of legitimate governance vary with the power exercised, the stakes of the decision, and the availability of alternatives.

Three considerations must be examined together. First, the institution's purposes must be defensible in light of the legitimate interests of affected parties. Second, the distribution and exercise of authority must be structured by appropriate legal and procedural constraints. Third, participants must possess effective ways to contest wrongful decisions and to influence rules that substantially structure their opportunities. These conditions do not require identical democratic participation in every product decision. They require a proportionate relationship between the public significance of authority and the public grounds on which that authority is exercised.

A platform's legitimacy is therefore relational. The same policy may be justified in one environment and indefensible in another because the dependence of participants and effects on third parties differ. A forum for a voluntary hobby group may use highly specialized membership rules; an infrastructurally indispensable system supporting public services has different obligations. A blanket requirement of neutrality can itself be unjust where it prevents legitimate protection against intimidation or discrimination. The standard is justified governance of the relevant environment, not the absence of governance.

Consent, Contract, and the Limits of Voluntariness

A central defense of platform authority invokes agreement. Users accept terms, sellers enter contracts, developers accept application policies, and workers choose whether to participate. Consent has genuine normative force when persons understand an arrangement, retain practical alternatives, and can revoke participation without disproportionate loss. It respects a domain of voluntary association in which people may establish specialized rules and purposes.

Consent also has conditions of meaningful exercise. Terms are frequently standardized, lengthy, revisable, and presented as prerequisites for access. The agreement may be formally valid without recording informed endorsement of every possible future change. Participants may join through public institutions or employment arrangements without selecting the underlying platform themselves. Others are affected without being users at all: a speaker's audience, a worker's family, a competitor, or a person profiled through another user's data can bear consequences without entering the relevant agreement.

The conceptual error would be to infer that private contractual form settles the political question. Even valid consent by the immediate parties does not automatically justify harms to third parties. Nor does it make every unilateral change in an ongoing relationship fair. Political philosophy has long understood that background institutions influence the opportunities within which formally voluntary choices occur. The fairness of a transaction may depend on prior allocations of resources, rights, legal standing, and available exit options.

A more adequate consent test asks whether the affected person could reasonably understand the relevant powers before agreeing, whether the promised terms are sufficiently stable, whether there is fair notice of important changes, and whether the person can exercise an effective right of objection. For relationships involving substantial dependence, legitimate arrangements may also require collective voice and independently enforceable standards. These requirements follow from the purpose of consent itself: a person should retain meaningful authority over decisions that bind that person.

A platform can meet these conditions imperfectly without losing all legitimacy. No institution can make every complex consequence foreseeable. The point is that an invocation of consent is a beginning of justification rather than a substitute for analyzing the power involved.

Freedom as Non-Domination and the Problem of Arbitrary Control

Republican freedom provides a particularly strong framework for assessing platform dependence. Philip Pettit's Republicanism articulates freedom as non-domination: the condition of not being subject to another's arbitrary capacity for interference. The ideal distinguishes actual intervention from the structure of power under which intervention remains possible. A dominant operator's discretion can affect a person's standing even where it chooses not to punish or exclude.

Applied to platforms, the relevant question is whether the operator's powers are appropriately constrained by the interests and claims of those subject to them. Participants may rely on an account, audience, payment channel, or technical ecosystem for years while remaining exposed to unpredictable suspension, alteration of terms, or discriminatory ranking. Their apparent liberty may depend on the goodwill or strategic calculation of an institution beyond their control. If no effective route of appeal, collective negotiation, independent oversight, or alternative participation exists, the relationship displays characteristics of domination.

A non-domination approach avoids the conclusion that moderation is inherently contrary to freedom. A clear and consistently enforced rule against targeted harassment may improve the conditions of participation for people previously intimidated into silence. An institution can interfere with harmful conduct in ways that protect equal freedom. The assessment turns on whether control serves defensible interests under accountable constraints, and whether the people subject to it can contest misuse.

There is an important difficulty. Pettit's theory was developed principally for political and social relations among persons. Corporate institutions possess organizational capacities that can be distributed among boards, managers, contractors, regulators, and technical systems. There need not be a single tyrannical individual whose will explains an arbitrary outcome. Institutional domination can arise through incentive structures that systematically deprive affected people of recourse. The absence of personal malice therefore does not establish the absence of arbitrary power.

Non-domination should be supplemented with attention to unequal capabilities. Exit may be affordable to well-resourced participants and devastating to others. A formal appeal can be meaningful to an institutionally sophisticated developer and inaccessible to a person without appropriate language support. The same platform policy can structure freedom differently across social positions. A normative assessment should therefore include both the availability of restraints on power and their practical accessibility.

Procedural Justice and the Right to Contest

Procedural fairness matters because platform decisions are repeated at scale under conditions of informational asymmetry. An affected person can encounter an adverse judgment without knowing the evidence used, the applicable rule, the available remedies, or the person responsible for correction. Such uncertainty produces more than inconvenience when access is connected with income, expression, organization, or reputation. It can deprive a participant of the practical means of defending legitimate interests.

A defensible framework distinguishes notice, explanation, hearing, review, and systemic accountability. Notice makes the applicable standard available. Explanation states why a decision was made in terms proportionate to its consequences. Hearing permits affected people to introduce relevant evidence or context. Review provides an effective route to correction. Systemic accountability examines whether the rule and its administration generate repeated patterns of error or unfairness. These protections vary according to urgency and risk; emergency restrictions responding to imminent danger may be imposed first and reviewed afterward.

Individual rights of appeal should be meaningful rather than ceremonial. An appeal mechanism fails its purpose if decisions are simply repeated by the same opaque process, if deadlines make redress useless, or if explanations are so generic that no relevant objection can be formulated. High-quality review may require trained personnel, clear allocation of responsibility, and authority to revise both a judgment and the policy that generated it.

Evelyn Douek's systemic account strengthens this analysis. A platform can achieve impressive statistics for prompt appeals while its ranking and moderation architecture systematically disadvantaged a population that rarely appeals or lacks awareness of the harm. Aggregate error rates can obscure concentrated costs. Institutions need both case-level routes of recourse and independent evaluation of outcomes, including attention to unequal burdens and structural incentives.

Oversight must itself be accountable. An external review board can become another remote authority if its appointment, mandate, evidence access, and relationship to affected communities are opaque. Judicial review and statutory regulation can be important, but courts may lack the expertise or capacity for continuous supervision of complex technical systems. A layered system of internal appeals, external auditing, regulatory supervision, public reporting, and legally enforceable rights may distribute corrective capacity more effectively than any single mechanism.

Democratic Authority and Private Administration

Democratic legitimacy raises the question of who should have a voice in establishing rules that shape common social life. A platform is not a territorial democracy: users do not necessarily form a single people with a uniform claim to electoral control, and different stakeholder groups can have conflicting interests. Nevertheless, platforms can regulate conditions relevant to democratic institutions. Their private status does not render the effects of their governance politically insignificant.

The normative insight of public-sphere theory is that participation in common reasoning depends on institutions that cannot be evaluated solely by their commercial efficiency. Habermas's account of publicity and Fraser's critique of exclusion show that the organization of public communication affects both the opportunities for criticism and the standing of different voices. A platform that shapes the visibility of public arguments becomes a participant in the institutional conditions of democratic life, even where it does not decide the substantive outcome of a vote.

Democratic accountability can take several forms. Public law can set baseline rights and obligations; independent regulators can supervise compliance; courts can address unlawful restrictions; civil society can conduct oversight; affected communities can participate in rule formation; and users can organize, exit, or support alternative networks. These mechanisms contribute different kinds of legitimacy. User voting cannot by itself settle legal rights, especially where a majority would suppress minority viewpoints. State rule-making cannot by itself guarantee pluralism, especially where political authorities seek to manipulate communication. Democratic design requires a structure in which multiple legitimate interests can be expressed and powers can check one another.

A public/private binary obscures the opportunity for institutional differentiation. Government may appropriately set minimum protections against discrimination, privacy violations, unfair market practices, or arbitrary exclusion while preserving space for private editorial judgment. Platforms may legitimately set community-specific standards within these limits. Independent institutions can verify compliance without taking over everyday moderation. The central normative question is where each authority should sit and which routes of challenge remain available when its decisions affect the public.

Freedom of Expression and the Editorial Rights of Platforms

Speech governance presents a conflict of legitimate claims. Individuals and associations have interests in expressing ideas, receiving information, and participating in public debate. Platform operators may have their own expressive interests in selecting, arranging, and presenting material. Other users have interests in safety, privacy, and freedom from targeted abuse. These claims cannot be settled simply by declaring that all content must remain available or that platform ownership confers complete discretion.

In Moody v. NetChoice, LLC (2024), the United States Supreme Court recognized constitutional protection for important forms of expressive curation undertaken by social media platforms and remanded the cases for proper analysis of the challenged laws' applications. The decision is a legal primary source for a particular constitutional framework. It should not be read as establishing identical protection for every technical function or answering every question about private domination. Nor does legal protection settle the philosophical evaluation of whether a platform's governance is fair to participants.

The analysis must distinguish expression from infrastructure. A newspaper's editorial selection is closely tied to its own expressive project. A hosting provider, payment processor, marketplace, and application store perform different functions. In practice, institutions can perform several simultaneously. Restrictions on one function may affect expressive freedom even when the function is not itself equivalent to publishing an editorial judgment. Philosophical clarity requires specifying the activity being regulated and the values that justify the regulation.

A state that compels a platform to display particular political messages may threaten freedom of expression. A platform that arbitrarily excludes vulnerable speakers may undermine their practical ability to participate. The appropriate response depends on the cause and structure of the harm. Greater transparency, narrowly targeted procedural obligations, nondiscrimination requirements, portability, interoperability, and public alternatives address different problems. Uniform carriage obligations may solve some access issues while undermining the ability of specialized communities to establish protective standards.

Free expression must be understood within a plurality of legitimate institutions. Public speech benefits from independent venues with varied editorial purposes, but independence becomes precarious when one or two infrastructures determine the feasible range of those venues. Protecting pluralism therefore involves both freedom for platforms to organize communities and structural conditions under which participants can form and sustain alternative communities.

Equality, Recognition, and Distributive Effects

Platforms often present general terms uniformly to their users, but uniform application does not guarantee equal effects. A ranking policy may favor speakers with existing resources, a verification system may impose greater costs on people with unstable identity documents, and a commission structure may affect small businesses differently from large firms. Evaluating equality therefore requires attention to the starting positions, vulnerabilities, and capacities of different participant groups.

Fraser's theory of counterpublics helps identify the importance of alternative spaces of association. Equal participation may require the ability to develop views, identities, and claims within communities that are not continuously subject to norms established by already dominant groups. Content policies intended to prevent harm can nonetheless reproduce exclusions if their categories ignore social context. At the same time, effective protections against harassment can make participation possible for groups previously excluded by hostile practices.

Safiya Umoja Noble's investigation of search and discriminatory representations demonstrates the value of looking beyond formal neutrality to social effects. A search result, recommendation, or classification can shape social visibility and esteem. Yet the existence of unequal outputs does not by itself identify the correct remedial institution. Some inequalities reflect underlying social conditions; others result from design choices, incentives, or discriminatory rule enforcement. Corrective action depends on diagnosing their sources.

Distributive assessment concerns control over the benefits created by participation. Platforms can create economic value by coordinating markets, but participants also produce content, data, reputation, social relations, and labor. Fairness demands examination of how rules allocate returns, risks, access, and opportunities to contest their allocation. No single distributive principle applies mechanically to all platforms. Nevertheless, an operator's control over the procedure through which benefits and burdens are distributed generates a reason for heightened scrutiny where participants are substantially dependent.

Privacy, Information Integrity, and Public Reason

Political freedom includes the capacity to form judgments without being subject to unreasonable surveillance and covert behavioral manipulation. When communication, consumption, education, and work occur through digital intermediaries, the organization of information flows becomes part of institutional power. Privacy is not exhausted by secrecy. Nissenbaum's contextual integrity framework emphasizes whether information moves in ways appropriate to the social setting and relationships involved.

Platforms can use data to improve performance and public safety. The same data may support targeting, profiling, discriminatory classification, or the prediction of vulnerabilities. Political evaluation requires asking whether the use of information is consistent with the purpose for which it was collected, whether affected people can identify consequential data practices, and whether there are meaningful constraints on secondary use. Broad consent to an indistinct category of future data processing gives little guidance about legitimate purposes.

Public reason is also affected by the quality of informational mediation. Commercial incentives can reward highly engaging material without regard to its epistemic reliability. Excessively centralized moderation can, however, concentrate authority to define acceptable knowledge. A satisfactory approach protects institutional diversity, supports independent scrutiny, permits legitimate expertise to inform decisions, and requires the operator to explain the standards that produce public-facing classifications.

The key distinction is between epistemic support and epistemic domination. A platform may help people find reliable information through transparent and challengeable procedures. It may also impose classifications that participants cannot adequately investigate or contest. Knowledge-related power is politically significant when classifications affect a person's ability to participate in collective judgment, defend a position, or maintain social standing.

Institutional Responsibility Without a Fictional Single Decision Maker

Responsibility for platform outcomes is frequently dispersed among designers, executives, moderators, contractors, advertisers, regulators, users, and automated systems. A harmful ranking outcome may be produced by an objective selected by management, a model trained on historical data, an interface optimized for engagement, and external incentives imposed by market competition. Attributing everything to the algorithm obscures the organizational decisions that made the system consequential.

Responsibility should follow control, knowledge, foreseeability, and capacity to remedy. The organization that selects objectives and maintains an infrastructure has responsibilities that cannot be transferred entirely to users merely because users generate content. A regulator that mandates a restriction bears responsibility for the legal justification of that mandate. A participating organization that uses a platform for sensitive public functions has obligations concerning procurement, oversight, and access. Different actors may share responsibility for different aspects of a system.

Automated decision making introduces a useful distinction between causal participation and normative accountability. An algorithmic process can be causally indispensable to a harmful outcome without possessing the institutional standing needed to provide justification or compensation. The platform operator remains responsible for its deployment choices, documentation, monitoring, and correction. Whether more sophisticated artificial systems eventually occupy different forms of legal or political standing is a separate philosophical inquiry. Platform responsibility can be analyzed fully without treating current algorithms as sovereign persons.

This distribution of responsibility suggests an institutional principle: accountability should be located where relevant decisions can actually be explained, reviewed, and changed. A complaint mechanism that lacks authority to alter a design objective cannot adequately address a systemic design problem. A legislative rule that delegates discretionary judgments to private operators should specify the nature and limits of the delegation. The allocation of accountability should map the actual allocation of control.

A Proportional Test for Justified Platform Governance

The preceding analysis supports a structured test. Its first element is the identification of the affected interests: speech, association, work, privacy, safety, equality, access, or other legitimate goods. Its second element is the operator's concrete governing capacity: exclusion, ranking, data processing, contracting, pricing, or infrastructural restriction. The third element is dependence: whether those affected have credible alternatives and the resources to exercise them. The fourth is justification: what legitimate purpose the intervention serves and why that intervention is suited to its purpose. The fifth is contestability: what constraints, oversight, rights of challenge, and mechanisms for correction exist.

These elements are not mechanically additive. A grave rights violation cannot be justified simply because alternative services exist. An urgent safety measure may be appropriate before individualized consultation, provided timely review follows. A highly competitive market may limit economic exploitation while leaving serious privacy or discrimination concerns. An institution with excellent procedures may still pursue a purpose fundamentally inconsistent with equal citizenship. Judgment requires attention to both substantive rights and the institutional form of their protection.

This test yields a general normative proposition: the more extensively a platform structures socially important opportunities, and the more dependent participants become on its infrastructure, the stronger the justificatory and accountability obligations attaching to its exercise of power. Those obligations should be proportionate to the relevant activity. They need not collapse all platforms into one public-law category.

Legitimate governance is therefore an ongoing relationship of justification, correction, and institutional restraint. It cannot be secured permanently by a declaration of neutrality, one initial contract, or one disclosure of design principles. Platform power changes as architectures, markets, social practices, and the capacities of the affected communities change. A responsible institutional order must be able to learn from these changes while preserving basic claims of freedom and equality.

Objections to the Political Interpretation of Platforms

The thesis that platforms can exercise political power requires careful qualification. A serious objection holds that digital platforms remain voluntary, privately owned organizations competing to provide valuable services; treating their ordinary management as political governance threatens to obscure the distinctive authority of the state. A second objection maintains that users can refuse the service or switch to alternatives, making domination an exaggerated description. A third argues that intervention in platform design or moderation risks converting private gatekeepers into instruments of governmental censorship. These objections identify real differences and dangers. Their force depends on whether the political interpretation claims identity with sovereignty, denies the significance of markets, or licenses unconstrained state power.

Robert Gorwa (2019) argues that platform governance must be understood through the relations among platform firms, users, states, and other institutions rather than reduced to internal company policy. Nicolas Suzor (2019) similarly analyzes the opaque private rules that govern digital life and the possibilities for principled constraints upon those rules. Their approaches support the framing adopted here: the object is a distributed field of governance whose powers remain differentiated across actors.\n\nThe theory defended here makes none of these claims. Political significance is a property of the effects and structure of a social relation, not merely of an institution's legal label. A private employer can affect freedom and status without becoming a state, and a privately controlled road or communication infrastructure can have significant public consequences without conferring general legislative authority on its owner. The same reasoning applies when a platform mediates important forms of association, expression, employment, or knowledge. Its powers may be politically salient while remaining legally and institutionally distinct from state sovereignty.

The objection from voluntary exchange points to a legitimate source of normative authority. People generally should be able to create communities, select services, enter contracts, and withdraw from associations. Yet the philosophical validity of consent depends in part upon the practical circumstances of choice. Network effects, sunk investments, nonportable reputations, and restricted access to alternatives may weaken the practical ability to leave. A service can be formally optional and still constitute a necessary route to an established audience, profession, or public institution. This does not prove domination in every case; it identifies the evidence needed to decide whether a claim of voluntariness does sufficient justificatory work.

The objection from state coercion is especially important. Governments possess powers of punishment and compulsory law that generally exceed those of private digital firms. Public authorities may also use legal obligations or informal threats to induce platforms to restrict political expression. The possibility of abusive state involvement limits the legitimate design of regulatory remedies. A political account of platform power should therefore examine both governmental and private concentrations of discretion, and specify safeguards for rights, institutional independence, and contestability.

The Case for Platform Autonomy

Platform autonomy protects valuable activities. A cooperative forum may choose to preserve a particular culture, a scientific journal may reject submissions inconsistent with its standards, and a digital service may organize information to suit an audience. General demands for viewpoint-neutral hosting would sometimes defeat the very forms of association and editorial judgment that make these institutions valuable. The right to create and govern a community can be an expression of freedom rather than an infringement of it.

The distinction between editorial freedom and infrastructural gatekeeping matters here. An editorial institution selects material as part of its communicative identity. A technical intermediary may exercise power through payment, hosting, identification, compatibility, ranking, or access to an underlying network. Some platforms combine these functions, so an appropriate account cannot classify an entire company under one normative heading. It must ask which activity is at issue, how much dependence that activity creates, and whose rights are implicated.

A strong argument for autonomy has three premises. First, freedom of association includes the ability to organize the purposes and standards of a voluntary institution. Second, pluralism depends on there being many institutions with different practices and editorial commitments. Third, an outside authority empowered to mandate the internal norms of all such institutions risks eliminating that pluralism. Therefore, intervention should preserve space for differentiated community governance. The conclusion is compelling, but its premises do not guarantee that the existing ownership structure provides a sufficiently diverse or contestable institutional environment. The very possibility of autonomous communities can depend upon access to a small number of upstream providers.

Protecting platform autonomy and restraining platform domination may therefore support one another. Interoperability and credible alternative infrastructures can make differentiated editorial communities more viable. Fair procedural standards can constrain arbitrary exclusion without compelling a particular substantive political viewpoint. Legal protections against covert state pressure can protect independent moderation. The aim is not to abolish platform choice but to distribute the conditions under which meaningful choice can be sustained.

Is Market Competition a Sufficient Remedy?

Competition can constrain pricing and service quality and can permit participants to refuse oppressive rules. A market with low switching costs, multiple providers, interoperable networks, and portable records may significantly reduce dependence on any one intermediary. This offers a powerful institutional alternative to direct regulation of every platform decision. The philosophical case for competitive discipline is that a credible right of exit can make unilateral power less arbitrary by exposing it to the practical possibility of rejection.

The effectiveness of this remedy is empirical and relational. Intermediaries operating multi-sided markets can benefit from network effects and economies of scale; participants can face the loss of contacts, ratings, accumulated audience relationships, or access to complementary technical services. Even where a competitor exists, the relevant choice may be between remaining under a disliked rule and abandoning valuable collective relations. An exit that requires forfeiting one's existing community is different from switching among fungible consumer products.

Competition also addresses only some types of injury. Several services can each collect disproportionate information or impose discriminatory standards. The possibility of using another service after wrongful suspension may offer little remedy for a person who has already lost income or suffered reputational damage. A market can furnish better choices while still producing systemic harms that fall on people outside the contractual relationship. Privacy, non-discrimination, and due process therefore retain distinct normative significance.

The appropriate conclusion is that competition is an indispensable constraint in many contexts and an incomplete theory of legitimacy. The institutional design question is which combination of competitive entry, technical portability, procedural rights, and public accountability addresses the specific powers and harms at issue.

Transparency, Explainability, and Their Limits

Transparency is a common proposed answer to opaque platform authority. Publishing terms of service, explaining moderation decisions, describing recommendation principles, and providing meaningful information about advertising or algorithmic ranking can improve understanding and permit criticism. Affected persons need enough information to know whether an intervention occurred and what grounds they might use to contest it. Independent researchers and public institutions also need appropriate access to evidence about aggregate effects.

Yet disclosure is not the same as effective control. A fully documented policy can remain unjust. An explanation can state the reason for an erroneous decision without providing authority to correct it. Technical descriptions may be unintelligible to ordinary participants or insufficient to reproduce an outcome. Transparency can also threaten privacy, expose confidential information, or create methods for evading legitimate safeguards. A policy of unrestricted source-code publication is therefore not a universal solution.

The distinction between intelligibility and contestability is central. An institution can be intelligible but entirely unilateral; another can be subject to powerful independent review while using some proprietary methods. The ideal is an accountable structure with appropriate access for affected parties, qualified auditors, competent regulators, and judicial authorities. Different kinds of explanation may be needed for a wrongful individual sanction, a discriminatory population-level pattern, and a public debate about the legitimate purpose of an algorithmic ranking system.

Evaluation must therefore ask who can act on the disclosed information. Public reporting is useful if civil organizations can challenge findings, if independent institutions can investigate, and if the platform can be required to remedy demonstrable problems. Otherwise transparency risks becoming a language of accountability detached from any effective distribution of corrective power.

Interoperability, Portability, and Institutional Pluralism

Interoperability allows independent systems to communicate under common technical conditions; portability allows persons or organizations to transfer certain data, relationships, or records. These capacities may strengthen exit and foster a plurality of communication and commercial institutions. Their philosophical significance lies in reducing the degree to which an individual's established social relations can be held within one operator's technical enclosure.

The normative promise is substantial. A creator who can preserve followers across services is less dependent on the arbitrary decisions of one platform. A business able to transfer legitimate reputational records has more bargaining power. Open communication protocols can permit communities to establish different rules while retaining opportunities for contact. Decentralized and federated systems may distribute the power to define social environments among a greater number of organizations and groups.

The promise has limits. Interoperability can create security risks when poorly specified interfaces allow abusive access, fraud, or data leakage. Conflicting moderation policies can make cross-platform communication difficult. Portability can expose information about persons who never consented to its transfer. A fragmented network can reproduce forms of dependence through hosting, domain infrastructure, application stores, or dominant protocol implementations. Institutional pluralism depends on effective governance of common standards, not on a simple change in software architecture.

The philosophical conclusion is that decentralized technologies provide opportunities for contestable governance whose realization depends on supporting institutions. A federated architecture can reduce some concentrations of power while generating new disputes over protocol stewardship, shared abuse prevention, and the rights of communities to set distinct terms. The desirability of pluralism requires the continued possibility of disagreement and exit under fair conditions.

Public Utility, Common Carrier, and Digital Constitutionalism

Analogies with public utilities and common carriers arise when an intermediary becomes widely relied upon for essential activities. The analogies suggest that certain forms of access should be provided under standardized, non-discriminatory conditions, sometimes with heightened oversight. They raise a fruitful question: when should private control over infrastructure be subject to duties that reflect its significance to the public?

The answer depends on the function. Mandatory carriage can be appropriate for some neutral transmission facilities yet objectionable if applied indiscriminately to expressive curation. A utility model may address dependence on basic infrastructure while failing to resolve the special problems of contextual moderation, search ranking, or the management of interpersonal harassment. The analogy is therefore a conceptual tool for identifying obligations associated with bottleneck power, rather than a ready-made classification of every digital platform.

Digital constitutionalism seeks principles and institutional arrangements capable of constraining power within the digital environment. The analogy to constitutional government is most useful when it directs attention to rights, justifications, distribution of functions, review, and the limits of discretionary power. It is misleading if a private terms-of-service document is treated as possessing the same legal origin or democratic authority as a public constitution. A private oversight body can contribute procedural independence without becoming a constitutional court.

The central design problem is institutional differentiation. Some decisions should be made by democratically accountable legislatures, others by independent courts or regulators, others by technical specialists, and still others by communities that create and maintain their own spaces. The allocation should reflect the kind of authority at stake, the affected interests, and the risks of institutional capture. Digital constitutionalism becomes a substantive research program when it specifies those allocations and the mechanisms through which people can challenge them.

Regulatory Examples: Digital Services and Digital Markets

The European Union's Digital Services Act, Regulation (EU) 2022/2065, is a contemporary legal response to the governance of intermediary services, including transparency, duties of care, notice-and-action procedures, and special obligations for very large online platforms and search engines. Its significance for philosophy lies in its attempt to distinguish categories of service and to relate regulatory responsibilities to systemic risks rather than simply declaring all intermediaries public authorities. Particular measures must still be evaluated for effectiveness, proportionality, and their effects on speech and privacy.

The Digital Markets Act, Regulation (EU) 2022/1925, addresses certain structural conditions of digital markets through obligations associated with designated gatekeepers. Its theoretical focus differs from that of the Digital Services Act. The former deals especially with contestability and fairness in the operation of core platform services; the latter addresses service governance and systemic risks. These separate objectives mirror the philosophical distinction between market dependence and the procedural or rights-related legitimacy of decisions over communication and information.

The regulations provide examples of institutional choices rather than conclusive proofs of normative success. A law can express a principled objective yet be poorly implemented, overbroad, or vulnerable to strategic compliance. A policy that improves user recourse may impose compliance costs that disadvantage smaller competitors. Demands for access to research data can support public knowledge while creating privacy and security risks. These tradeoffs require ongoing empirical scrutiny and public justification.

In the United States, Moody v. NetChoice, LLC (2024) shows a different institutional context in which constitutional protection for expressive curation constrains how governments may regulate platform moderation. It also illustrates why comparative analysis must distinguish legal traditions and specific platform functions. Neither an American constitutional decision nor European legislative architecture establishes a universal philosophy of platform power. Each reveals the continuing task of allocating responsibility among private actors, public law, and the people who depend upon digital infrastructures.

Critical Theory and the Problem of Reproduction

A deeper objection holds that procedural reforms leave intact the political economy that produces platform dependence. From this perspective, independent appeals, transparency reports, and rights to notice can legitimize a system whose economic incentives continue to favor data extraction, concentration, and precarious labor. Srnicek's political economy and Cohen's legal-institutional analysis direct attention to the structural production of platform positions. Zuboff emphasizes practices in which behavioral information becomes a strategic economic asset. The argument is that fairness at the level of individual decisions cannot fully address asymmetry at the level of ownership, investment, and accumulation.

This challenge has considerable force. A platform can administer its rules consistently while allocating the benefits of participation in ways that remain unjustifiable. It can publish transparent policies while using its infrastructural position to impose disadvantageous terms. Institutional reforms therefore must include the background structure of markets and the legal powers through which platforms acquire and maintain their positions.

Yet an exclusively structural critique risks overlooking the distinct significance of procedural injuries. Workers and users require immediate remedies even when the wider political economy cannot be transformed quickly. The availability of reasons, appeal, evidence, and independent investigation can also expose structural patterns and enable further collective action. The approaches are complementary when procedural rights are treated as instruments of accountability and structural reforms as changes to the conditions under which power is reproduced.

The assessment of platform power must thus operate at different scales: immediate decisions, organizational procedures, market structure, and the legal-political constitution of infrastructural control. A reform that improves one scale can be valuable while leaving problems at the others unresolved.

The Limits of Collective Governance

User councils, platform cooperatives, multistakeholder boards, collective bargaining, and community voting promise to give participants a role in governing the systems they use. Such mechanisms can improve legitimacy by introducing experience and claims unavailable to managers alone. They also face questions about representation, incentives, expertise, and unequal participation. A single category of "users" may conceal conflicting interests among workers, creators, advertisers, consumers, vulnerable speakers, and persons affected without holding accounts.

Democratic participation is therefore not simply a matter of collecting votes. The governance institution needs justified procedures for defining who is affected, determining which interests are protected by rights, establishing representation, and resolving disputes. Majority decisions cannot automatically justify discriminatory exclusion, invasions of privacy, or arbitrary deprivation of livelihood. Representative procedures may protect against managerial discretion while introducing new concentrations of power among highly organized participants.

Institutional experimentation remains important. Different services may benefit from cooperative ownership, elected advisory councils, independent appeals, worker representation, public oversight, or open protocols. A generalized requirement to adopt one internal constitutional structure would overlook differences in scale and purpose. The normative criterion should be the effective accountability of consequential power and the ability of participants to form alternative institutions.

One further distinction is decisive: consultation is different from shared authority. An operator may invite feedback while retaining the unrestricted right to ignore it. Genuine collective governance requires specified powers to obtain information, influence rules, require review, and secure compliance with decisions. Participation achieves political significance when it changes the distribution of institutional capacities.

Artificial Systems and the Historical Expansion of Platform Relations

Contemporary platforms increasingly mediate interactions involving generative artificial intelligence, automated agents, software services, and computational systems that act in concert with institutions and persons. This development expands the categories of activities governed by platform rules. Access policies may determine which artificial systems can operate, which tools they can call, what information they can retrieve, and whether they can interact with users or other systems. The platform becomes an organizer of relations whose participants and causal processes may extend beyond direct person-to-person communication.

The philosophical significance can be stated without importing contested claims about the consciousness or moral status of artificial systems. Platform governance can shape human interests through automated activity, and organizations remain responsible for the systems they deploy. Questions about whether an artificial system can possess independent legal standing or public agency belong to distinct inquiries. The platform-power problem concerns who has authority to organize and constrain mediated interactions and what justificatory duties follow.

Within the horizon called Artificial Era in the Aisentica research program, the transformation of digital infrastructure contributes to the larger question of how different forms of agency, reasoning, and institutional participation are ordered. The present monograph uses that horizon as a conceptual context and preserves the distinction between its original philosophical theses and the independent evidence examined in political philosophy, platform studies, and law. Its findings about platform power do not establish an ontological theory of Artificial or a claim about any particular system's sapience.

The coming philosophical questions concern layered governance in hybrid environments, the accountability of delegated decisions, the portability of interaction histories, and the extent to which shared infrastructures should enable plural forms of human and artificial participation. The same principles developed above continue to apply: consequential power requires identified purposes, proportionate constraints, effective remedies, and institutions capable of challenging arbitrary authority.

Comparative Philosophical Findings

The historical and conceptual analysis yields a structural account of platform power. Classical theories illuminate different aspects of a contemporary phenomenon: Hobbes connects authority with collective order and authorization; Weber describes organized domination through regularized administration; Arendt reveals the distinctive power of people acting in concert; Foucault explains the government of conduct through structured possibilities; and Pettit clarifies why exposure to arbitrary interference can compromise freedom before any punitive decision takes place. None of these theories is reducible to the others. Their combined relevance lies in showing that power concerns capacities, institutions, practices, relationships, justifications, and the conditions in which people act.

The philosophy of technology supplies a second layer. Winner demonstrates that artifacts can embody consequential political choices, while Lessig explains how code can regulate conduct alongside law, markets, and social norms. Platform studies adds the distinctive organizational form: an intermediary controls the technical and contractual architecture of participation, collects information about mediated activities, and can revise the conditions of connection. Theories of informational capitalism explain how these capacities are encouraged and stabilized by legal institutions, economic incentives, and accumulated dependence. The digital public sphere shows why their consequences extend to collective knowledge, association, and political speech.

The analysis also identifies a conceptual sequence. A platform first enables an interaction and establishes conditions of use. Its relative importance increases as participants develop relationships and practices dependent upon those conditions. The operator's ability to change rules, ranking, access, information practices, or compensation becomes more consequential as the costs of departure increase. The power becomes politically significant when these decisions materially affect interests that depend on shared institutions and recognized claims to freedom, equality, livelihood, and public participation.

These findings distinguish the political significance of platform power from the legal identity of the platform operator. A private corporation can affect political conditions without becoming a sovereign government. A public agency can rely on private infrastructure without transferring all public responsibilities to it. An algorithm can enact an institutional decision without independently possessing the normative authority of the institution. A contract can govern a voluntary relationship without settling the legitimacy of every subsequent rule or every effect on third parties. Each distinction prevents the analysis from collapsing into a misleading analogy.

The normative synthesis is correspondingly plural. Freedom of association matters because people should be able to create distinct communities with their own legitimate purposes. Freedom from domination matters because participants should not depend upon unrestrained discretion where essential interests are at stake. Fair procedures matter because errors and exclusion must be open to correction. Structural plurality matters because exit is valuable only when credible alternatives can preserve important relationships and opportunities. Public accountability matters because privately administered infrastructures can materially shape the conditions of democratic life. A theory of legitimate platform power must hold these considerations together.

Open Questions for Continuing Research

Several problems remain unresolved at the philosophical and institutional levels. First, what threshold of dependence makes a platform a public-interest infrastructure? Size, market share, functional indispensability, network centrality, and the severity of exclusion offer different indicators. A single threshold may be inadequate because local communities, specialized industries, and geographically concentrated services can experience severe dependence without satisfying broad numerical criteria. Future work should examine multidimensional standards that combine the practical costs of loss of access with the affected interests.

Second, how should platform constituencies be identified? Some people use the service directly; others are affected by recommendations, profiling, labor conditions, procurement, or upstream infrastructure without being contractual users. A democratic or participatory governance model cannot assume an obvious, bounded membership. It must specify who has standing to challenge a rule and what form of representation is appropriate to different effects.

Third, how should legal and political institutions divide their supervisory responsibilities? Courts can protect rights and correct particular violations. Regulators can evaluate systemic patterns and enforce standards. Competition authorities can address gatekeeper conduct and structural barriers. Collective institutions can supply experiential knowledge and bargaining power. Independent auditors can test technical claims. The question is how these forms of oversight should be combined without generating excessive centralization or conflicting obligations.

Fourth, how do changing technical arrangements alter effective power? Federation, portability, privacy-preserving protocols, user-controlled recommender systems, and decentralized identity may shift particular capacities. Their success depends on governance of technical standards, security, incentives, and accessibility. A system that disperses ownership may still concentrate crucial protocol decisions. Evaluation should follow the actual ability of actors to set, challenge, and revise the conditions of participation.

Fifth, the increasing intermediation of activities involving artificial systems raises questions about representation, delegation, and the attribution of actions across human and nonbiological infrastructures. These questions deserve separate investigation within philosophy of technology, action, law, and political theory. The present results establish a method for evaluating platform authority over mediated relations; they do not determine the ontological or legal status of the systems participating in those relations.

Finally, the inquiry must remain empirical as well as philosophical. Whether a regulation reduces wrongful removal, whether an appeals process corrects systematic bias, whether interoperability strengthens actual exit, and whether collective governance distributes effective authority are questions requiring evidence. Normative principles determine what should be assessed; empirical investigation determines which institutional arrangements satisfy those principles in concrete settings.

Conclusion

Platform power is an institutional capacity to govern the conditions of mediated participation. It operates through rule-making, visibility, information, market coordination, labor management, and control of infrastructural dependencies. Its exercise can create valuable possibilities for cooperation and communication; it can also expose participants to discretionary decisions that affect expression, livelihood, association, privacy, and democratic participation. The relation becomes politically significant when the intermediary controls important shared opportunities and those affected lack sufficiently effective means to contest or refuse its terms.

The monograph has established that platform authority is historically continuous with problems of political power while possessing distinctive contemporary mechanisms. Classical theories of sovereignty and authorization clarify why the capacity to make rules requires justification. Institutional sociology explains the organization of compliance. Republican theory identifies subjection to arbitrary power as a matter of freedom even in the absence of actual sanction. Theories of technological mediation explain how rules can be embodied in the architecture of action. Political economy and infrastructure studies show how legal structures, market concentration, and accumulated dependence can stabilize governing positions.

The resulting philosophical position is that legitimacy depends on more than ownership, contractual acceptance, useful technical performance, or transparency. It requires a defensible relationship between an operator's purposes, the concrete powers it exercises, the interests affected by its decisions, and the institutional means available for correction. Platforms with different functions and degrees of public consequence warrant differentiated obligations. Rights-protective procedure, independent evaluation, meaningful exit, interoperability where suitable, collective representation, competition, and public law are complementary instruments whose proper combination must be established for each setting.

The deepest consequence concerns the structure of freedom in mediated societies. Freedom includes the ability to act within a digital environment and the institutional capacity to question the rules that constitute that environment. Individuals and communities may be offered abundant opportunities for action while remaining dependent upon the unilateral discretion of those who organize access. Legitimate platform governance enlarges the range of practical possibilities while placing consequential authority under justifiable constraints.

The philosophy of platform power accordingly becomes a philosophy of the institutional conditions of participation. Its central question is who may organize the possibilities through which others communicate, work, know, associate, and act together, and on what grounds that organizing authority may be exercised. The answer is a regime of justified, accountable, proportionate, and contestable governance, sustained by the effective pluralism of institutions and the ability of affected people to exercise counterpower. The same question acquires wider historical scope as digital infrastructures mediate relations among humans, organizations, and artificial systems in the Artificial Era.