1. Introduction

Market power and epistemic democracy risks concern the possibility that economically dominant firms acquire disproportionate control over the production, distribution, visibility and validation of information, thereby weakening the conditions necessary for informed democratic participation.

Traditional competition law examines whether a firm can profitably raise prices, restrict output, exclude rivals or exploit customers. In digital markets, however, a firm may exercise significant power even when its services are free. Search engines, social networks, app stores, online marketplaces, cloud providers and artificial intelligence (AI) systems can influence what information people encounter, which speakers reach audiences, which businesses become visible and which sources are treated as credible.

Epistemic democracy refers to the idea that democratic decision-making depends partly on citizens having access to diverse, reliable, contestable information and opportunities to challenge prevailing claims. It does not require citizens to agree on the truth of every proposition, nor does it permit governments to impose a single officially approved worldview.

The competition-law concern is that concentrated economic power can become concentrated informational power. Where a small number of firms control essential information intermediaries, the same commercial incentives that shape markets may also affect the quality and diversity of public discourse.

The legal challenge is to distinguish ordinary editorial discretion and legitimate product design from conduct that entrenches market power, forecloses competitors or undermines meaningful user choice.

2. Meaning and principal concepts

A. Market power

Market power is the ability of an undertaking to act to an appreciable extent independently of competitive constraints. It may arise from:

Network effects that make a platform more valuable as more people use it.

Economies of scale and scope in data collection, computing and distribution.

Control over default settings, search rankings, app distribution or advertising infrastructure.

High switching costs, weak interoperability and limited data portability.

Vertical integration between the platform and businesses competing on it.

Control over foundation models, cloud computing, identity systems or other critical AI infrastructure.

Dominance is not itself unlawful under Article 102 TFEU or Chapter II of the UK Competition Act 1998. The concern is abusive conduct, subject to the applicable legal test.

B. Epistemic power

Epistemic power is the capacity to influence what information is available, discoverable, credible or practically usable. It may operate through:

Selection: determining which information appears in search results or recommendation feeds.

Visibility: allocating attention through ranking, recommendation and advertising systems.

Classification: labelling content as trustworthy, misleading, relevant or low quality.

Aggregation: combining large datasets to produce insights unavailable to competitors.

Validation: using platform signals, ratings or AI-generated summaries to influence perceived credibility.

Infrastructure: controlling the systems through which journalists, researchers, publishers and citizens reach their audiences.

A firm may possess epistemic influence without having a legal monopoly over information. Conversely, a large audience alone does not establish competition-law dominance.

C. Epistemic democracy risk

The risk emerges when concentrated informational control weakens the conditions for public reasoning. Examples include:

Independent publishers becoming economically dependent on a dominant search or social platform.

Algorithmic ranking systematically disadvantaging competing sources or viewpoints for commercially strategic reasons.

AI assistants directing users towards a narrow set of sources while obscuring uncertainty or alternative evidence.

A dominant platform changing access rules in ways that undermine the viability of independent news, research or public-interest services.

Data and advertising concentration making it difficult for new entrants to challenge incumbent information intermediaries.

These are possible mechanisms of harm, not automatic findings of illegality. Evidence is required to establish the relevant market, competitive effects and legal basis for intervention.

3. Legal framework

A. European Union law

Article 102 TFEU prohibits abuse of a dominant position within the internal market where the relevant conditions are satisfied. It may apply to exclusionary practices that restrict rivals' access to users, distribution channels, data or other commercially important inputs.

Article 101 TFEU prohibits agreements and concerted practices that restrict competition, including certain arrangements that divide markets, limit independent information services or coordinate competitive conduct.

The Digital Markets Act (DMA) imposes specified obligations on designated gatekeepers. These obligations can address conduct such as self-preferencing, restrictions on business users and barriers to interoperability, subject to the applicable provisions. Its framework is distinct from traditional dominance-based enforcement.

The EU Charter of Fundamental Rights also supplies relevant constitutional context:

Article 11 protects freedom of expression and information, including the freedom to receive and impart information.

Article 16 protects the freedom to conduct a business.

Article 47 protects the right to an effective remedy and a fair hearing.

These rights must be reconciled. Competition authorities cannot simply equate editorial choices with abuse, nor can platforms automatically invoke editorial freedom to immunise exclusionary commercial conduct.

B. German competition law

The principal provisions of the German Act Against Restraints of Competition (Gesetz gegen Wettbewerbsbeschränkungen, or GWB) include:

Section 19 GWB: prohibits abusive conduct by dominant undertakings.

Section 19a GWB: enables the Bundeskartellamt to impose specified measures on undertakings of paramount significance for competition across markets, subject to the statutory requirements.

Section 20 GWB: addresses specified conduct by undertakings with relative or superior market power and certain other circumstances covered by the provision.

Section 19a is particularly relevant where a platform's position across multiple markets enables it to transfer advantages between search, advertising, social networking, app distribution, cloud infrastructure or AI services.

German competition law can therefore address certain forms of cross-market dependency before they produce the same effects as traditional price-based exclusion. However, a general reduction in public discourse quality is not, by itself, sufficient to establish a violation of these provisions.

C. United Kingdom law

The UK framework includes:

Competition Act 1998, Chapter II: prohibits abuse of a dominant position.

Digital Markets, Competition and Consumers Act 2024: establishes a regime for designated firms with strategic market status and provides for tailored conduct requirements and pro-competition interventions.

Human Rights Act 1998 and Article 10 ECHR: provide relevant protections for freedom of expression, subject to applicable limitations.

The UK regime can address the commercial structures that shape access to digital audiences. Nevertheless, competition-law remedies must rest on the relevant statutory powers, evidence and procedural safeguards.

4. Principal mechanisms linking market power to epistemic democracy risks

1. Search and recommendation gatekeeping

Ranking systems determine which sources users are likely to discover. If a dominant intermediary favours its own commercial services or disadvantages competing services, the effect may extend beyond lost sales to reduced visibility for independent information providers.

2. Algorithmic concentration of attention

Personalised feeds can concentrate audience attention around a limited set of sources. Competition law becomes relevant when exclusionary platform practices, rather than personalisation alone, weaken the ability of competing services to reach users.

3. AI-mediated knowledge dependence

Users may increasingly rely on a small number of AI assistants for summaries, explanations and recommendations. Control over models, training data, inference infrastructure and distribution can create entry barriers and make alternative information services less viable.

4. Economic dependence of publishers

Publishers may depend on dominant intermediaries for traffic, advertising revenue or audience analytics. Sudden changes in access, ranking or monetisation can undermine independent providers even where no explicit agreement to suppress information exists.

5. Infrastructure and data concentration

Exclusive access to valuable datasets, cloud capacity, identity systems or model-serving infrastructure can make independent research and alternative information products more expensive or technically impracticable.

5. At least 8 important case laws

The following cases establish relevant competition-law principles. They do not all decide epistemic democracy as an independent legal doctrine. Their significance lies in how their holdings can inform the analysis of informational gatekeeping, market access, data concentration and platform power.

Case 1: Google and Alphabet v Commission (Google Shopping)

Citation: Case C-48/22 P, Court of Justice of the European Union, 10 September 2024.

Facts and issue: The European Commission found that Google had favoured its own comparison-shopping service in general search results while disadvantaging competing comparison-shopping services.

Legal principle: The Court of Justice dismissed Google's appeal, upholding the judgment that sustained the Commission's finding of abuse. The case confirms that conduct by a dominant platform which favours its own downstream service can constitute abuse where the applicable conditions for anticompetitive foreclosure are established.

Relevance to epistemic democracy: Search rankings allocate attention. Preferential treatment can make independent services less discoverable, potentially narrowing the range of sources through which users obtain information. The legal finding, however, concerned competition in comparison shopping, not political bias or democratic harm as such.

Case 2: Google and Alphabet v Commission (Google Android)

Citation: Case T-604/18, General Court of the European Union, 14 September 2022.

Facts and issue: The case concerned Google's Android-related practices, including contractual restrictions associated with mobile application distribution and search.

Legal principle: The General Court largely upheld the Commission's finding of abuse while adjusting the fine. The judgment illustrates how contractual arrangements, defaults and distribution conditions may reinforce the position of a dominant digital ecosystem.

Relevance to epistemic democracy: Default search arrangements and pre-installation can affect which information services users encounter. When users face friction in switching, commercial distribution advantages may become durable informational advantages.

Case 3: Microsoft v Commission

Citation: Case T-201/04, Court of First Instance of the European Communities, 17 September 2007.

Facts and issue: The case concerned Microsoft's dominance in PC operating systems, interoperability information and the tying of Windows Media Player to Windows.

Legal principle: The Court largely upheld the Commission's decision, including findings concerning refusal to supply interoperability information and tying.

Relevance to epistemic democracy: Interoperability restrictions can make it harder for rival developers to build services that operate effectively alongside a dominant platform. The principle is relevant to AI ecosystems where access to interfaces, system information or essential technical compatibility may determine whether alternative information services can compete.

The judgment does not create a general obligation to disclose all proprietary information; the legal conditions governing the particular form of abuse remain critical.

Case 4: Bronner v Mediaprint

Citation: Case C-7/97, Court of Justice of the European Communities, 26 November 1998.

Facts and issue: A small newspaper publisher sought access to a rival publisher's nationwide home-delivery distribution system.

Legal principle: The Court established a demanding test for imposing a duty on a dominant undertaking to grant access to infrastructure it has developed for its own business. The relevant conditions include indispensability and the risk of eliminating all competition by the undertaking seeking access, subject to the precise circumstances.

Relevance to epistemic democracy: If a distribution network becomes indispensable for reaching an audience, exclusion may have consequences for informational diversity. Yet Bronner cautions against assuming that every commercially important platform must share its infrastructure.

Case 5: Slovak Telekom v Commission

Citation: Case C-165/19 P, Court of Justice of the European Union, 25 March 2021.

Facts and issue: The case concerned a telecommunications operator's access conditions and margin-squeeze conduct affecting competing operators.

Legal principle: The Court clarified the relationship between refusal-to-supply doctrine and other forms of abusive conduct, including margin squeeze. The strict Bronner indispensability test does not automatically govern every form of access-related abuse.

Relevance to epistemic democracy: A platform may formally provide access while setting commercial terms that make effective competition difficult. In digital information markets, discriminatory access fees, technical limitations or vertically integrated pricing can weaken independent providers without an outright refusal of access.

Case 6: United Brands v Commission

Citation: Case 27/76, Court of Justice of the European Communities, 14 February 1978.

Facts and issue: The case concerned United Brands' position in the banana market and several alleged abuses of dominance.

Legal principle: The Court developed foundational principles on defining dominance and assessing whether an undertaking can behave to an appreciable extent independently of competitors, customers and consumers.

Relevance to epistemic democracy: The case supplies the underlying market-power framework. Applied to information intermediaries, the question is whether users, publishers and advertisers have realistic alternatives or whether network effects, switching costs and dependency allow the platform to act with limited competitive constraint.

It does not establish that economic dominance necessarily produces political or informational dominance.

Case 7: Hoffmann-La Roche v Commission

Citation: Case 85/76, Court of Justice of the European Communities, 13 February 1979.

Facts and issue: The case concerned dominance in vitamins and loyalty-inducing arrangements with customers.

Legal principle: The Court explained that dominance entails a position of economic strength and that certain loyalty arrangements by a dominant undertaking can restrict competition by tying customers to it and hindering rivals.

Relevance to epistemic democracy: Long-term dependency arrangements in digital ecosystems may make publishers, advertisers or developers reluctant to use alternative channels. The analogy is strongest where contractual incentives or restrictions reinforce lock-in; it should not be extended automatically to ordinary loyalty programmes or user preferences.

Case 8: Meta Platforms and Others

Citation: Case C-252/21, Court of Justice of the European Union, 4 July 2023.

Facts and issue: The case concerned the relationship between competition enforcement, personal-data processing and the German competition authority's proceedings against Meta.

Legal principle: The Court addressed the circumstances in which a national competition authority may examine compliance with the GDPR when assessing an abuse-of-dominance theory. It also emphasised the need for cooperation with the relevant data-protection supervisory authorities and for distinguishing data-protection questions from competition-law conclusions.

Relevance to epistemic democracy: Personal data can support highly effective targeting, personalisation and inference. Concentration of data may strengthen a platform's competitive position and its capacity to shape information environments. The judgment helps explain how competition law and data-protection law can interact without treating every GDPR infringement as automatically constituting an antitrust violation.

Case 9: Bundeskartellamt v Facebook (Meta)

Citation: German Federal Court of Justice, Case KVR 69/19, 23 June 2020.

Facts and issue: The Bundeskartellamt challenged Facebook's collection and combination of user data from Facebook and third-party services, finding an abuse linked to its dominant position and the terms imposed on users.

Legal principle: The Federal Court of Justice upheld the central basis for the authority's intervention at the interim stage. The proceedings illustrated that exploitative data practices may be relevant to competition law when linked to market power and the applicable legal requirements.

Relevance to epistemic democracy: Extensive cross-service data aggregation may reinforce a platform's capacity to personalise content, target audiences and predict behaviour. This may create informational asymmetries and make entry by rival services more difficult.

Important qualification: The case does not establish that personalised advertising, extensive data collection or the exercise of editorial discretion is inherently unlawful.

Case 10: Airtours v Commission

Citation: Case T-342/99, Court of First Instance of the European Communities, 6 June 2002.

Facts and issue: Airtours challenged the European Commission's prohibition of its proposed acquisition of First Choice, which was based on concerns about collective dominance in the UK package-holiday market.

Legal principle: The Court annulled the Commission's decision because the evidence did not sufficiently establish the conditions necessary for collective dominance.

Relevance to epistemic democracy: The case demonstrates the importance of evidence when assessing concentrated market structures. Claims that a merger will reduce informational diversity, diminish independent journalism or strengthen a platform's control over knowledge must be supported by a rigorous analysis of market conditions and likely effects.

The case is particularly useful as a warning against substituting general concerns about concentration for a demonstrated causal theory of harm.

6. How should epistemic democracy risks be assessed?

A competition authority assessing these issues should distinguish economic concentration, informational influence and legally cognisable competitive harm.

Stage 1 — Define the relevant market

Identify whether the relevant activity involves general search, social networking, digital advertising, app distribution, AI assistants, cloud computing or another service. Examine substitution, multi-sided interactions and non-price competition.

Stage 2 — Establish market power

Evaluate market shares, entry barriers, network effects, data advantages, switching costs, defaults and dependence on the platform.

Stage 3 — Identify the challenged conduct

Determine whether the conduct involves self-preferencing, discriminatory access, tying, exclusionary contracts, margin squeeze, data-related foreclosure or another recognised legal concern.

Stage 4 — Establish the causal mechanism

Test whether the conduct impairs rivals' ability to compete, limits entry, reduces innovation or restricts user choice. Informational diversity can be relevant evidence where it is connected to the applicable competition-law test.

Stage 5 — Evaluate remedies and safeguards

Consider whether interoperability, portability, non-discrimination, transparent access conditions or structural measures would address the proven harm. Protect confidential information, due process and legitimate editorial autonomy.

7. Economic analysis and measurement

Epistemic democracy risks are difficult to measure because information quality and diversity cannot be reduced to a single competition metric.

Several indicators may nevertheless assist an investigation.

A. Counterfactual analysis

An investigator should ask what would likely happen in the absence of the challenged conduct.

For example, if a search platform promotes its own service, the analysis might compare actual referral traffic and rival entry with a plausible scenario in which competing services receive non-discriminatory treatment.

The counterfactual must account for differences in product quality, relevance, consumer preferences and competing explanations for market outcomes.

B. Welfare beyond price

In zero-price services, consumer harm may appear through reduced quality, weaker privacy, less innovation or reduced choice rather than higher monetary prices.

However, a decline in source diversity should not automatically be treated as proof of consumer harm or abuse. The authority must connect the observed outcome to the relevant legal standard and demonstrate why the conduct is attributable to the undertaking.

8. Constitutional and democratic implications

The issue goes beyond ordinary commercial performance because information intermediaries can influence the practical conditions under which democratic societies debate public questions.

Three competing interests require careful balancing.

First, freedom of expression. Platforms and publishers have legitimate interests in deciding what content to host, recommend or remove. Competition remedies should not turn regulators into general arbiters of acceptable political or social opinion.

Second, economic freedom and innovation. Firms must retain incentives to develop new search technologies, recommendation systems, AI models and information products. Mandatory access or disclosure can undermine investment when designed without regard to cost, security and intellectual-property interests.

Third, pluralism and contestability. A market with meaningful entry, interoperable services and alternatives for users is less vulnerable to unilateral control by a single intermediary. Competition law can protect the competitive conditions that allow such alternatives to emerge, even though it cannot guarantee democratic consensus or eliminate misinformation.

The central constitutional distinction is between regulating the competitive structure through which information is distributed and dictating the information citizens should receive. The former may fall within competition-law powers; the latter raises substantially different legal and democratic concerns.

9. Remedies and policy responses

Potential responses should be tailored to the established source of harm.

Non-discrimination obligations: prevent a dominant intermediary from unfairly disadvantaging rival services where the applicable law permits such intervention.

Interoperability: reduce technical barriers that prevent users and businesses from adopting alternatives.

Data portability: facilitate switching and reduce dependence on accumulated user data, subject to privacy and security safeguards.

Transparent access conditions: improve the ability of publishers, developers and business users to understand and challenge commercial restrictions.

Merger scrutiny: examine whether acquisitions eliminate emerging competitors, potential rivals or important sources of innovation.

Independent audits: test ranking systems, access rules or algorithmic effects where legally authorised and proportionate.

Procedural safeguards: provide reasons, review mechanisms and opportunities to challenge consequential platform decisions.

No single remedy is appropriate in every case. For example, transparency alone may not solve foreclosure, while mandatory access may be unnecessary where users have effective alternatives.

10. Critical evaluation

The strongest argument for incorporating epistemic concerns into competition analysis is that information intermediaries compete on dimensions other than price. Their conduct may influence access, quality, innovation and the ability of independent providers to survive. A purely price-centred analysis can overlook some of these effects.

The principal objection is that epistemic democracy is not a free-standing prohibition under Article 102 TFEU, Chapter II of the UK Competition Act 1998 or the GWB. If the concept is used too broadly, it could encourage intervention based on disagreement with a platform's editorial decisions rather than demonstrated competition concerns.

A sound legal approach therefore requires:

A clear theory of competitive harm.

Evidence of market power or the statutory conditions for the relevant intervention.

A credible causal connection between the conduct and the alleged harm.

Consideration of efficiencies, innovation and legitimate product design.

Proportionate remedies and effective judicial review.

The ten cases discussed above provide tools for analysing dominance, exclusion, access, data practices and evidentiary standards. They do not establish that competition law guarantees informational pluralism as an independent outcome.

11. Conclusion

Market power can become epistemic power when a small number of undertakings control the infrastructure through which information is discovered, distributed, monetised and interpreted. Network effects, data advantages, default settings, algorithmic ranking and AI infrastructure can make this power durable and difficult to contest.

EU, German and UK competition law offer mechanisms for addressing some underlying causes, particularly exclusionary conduct, discriminatory access, tying, exploitative practices where legally established and structural barriers to entry. Cases such as Google Shopping, Microsoft, Bronner, Meta and Airtours help identify both the scope and the limits of intervention.