Competition Law and Intelligent Influence Infrastructures and Market Power

Introduction

Intelligent Influence Infrastructures (IIIs) may be understood as digital, algorithmic and data-driven systems that influence how consumers, businesses and market participants see, rank, select, recommend, purchase, access or interact with products and services.

They include:

• AI recommendation engines;
• search and ranking algorithms;
• personalised advertising systems;
• recommender systems;
• digital choice architecture;
• default settings;
• app-store ranking systems;
• marketplace Buy Boxes;
• algorithmic pricing and auction systems;
• social-media feeds;
• AI assistants and agents;
• interoperability and API-control systems; and
• data-driven personalisation infrastructures.

Competition law becomes relevant when an undertaking possessing substantial market power can use such infrastructure not merely to improve its product, but to shape competitive conditions in favour of itself or its affiliated businesses.

The central competition-law question is therefore:

When does intelligent influence cease to be ordinary competition by innovation and become an instrument for acquiring, maintaining or exploiting market power?

Modern competition authorities increasingly examine this issue through dominance, self-preferencing, tying, exclusionary conduct, discriminatory access, data advantages, interoperability restrictions, exploitative conduct and algorithmic coordination.

I. Meaning of Intelligent Influence Infrastructure

An intelligent influence infrastructure is more than an algorithm.

It is a combination of:

Data + algorithm + interface + ranking + recommendation + defaults + behavioural feedback + market access.

For example:

User searches → algorithm predicts preference → platform ranks products → interface highlights selected products → user chooses highlighted product → transaction generates new data → algorithm learns → future ranking changes.

This creates a feedback loop.

The infrastructure can therefore influence both:

1. consumer behaviour, and
2. competitive behaviour of rival firms.

The competition concern becomes particularly serious where the infrastructure is controlled by a firm that also competes with the businesses whose products are being ranked.

II. Legal Framework

1. Abuse of Dominant Position

The traditional starting point is whether the undertaking possesses substantial market power.

Relevant legal regimes include:

European Union

• Article 102 TFEU;
• Digital Markets Act (DMA);
• merger-control principles;
• competition rules concerning digital ecosystems.

United States

• Section 2 Sherman Act;
• Section 5 FTC Act;
• Clayton Act where acquisitions or structural transactions are involved.

United Kingdom

• Chapter II Competition Act 1998;
• Digital Markets, Competition and Consumers Act 2024.

India

• Competition Act 2002, particularly:
  • Section 4 — abuse of dominant position;
  • Section 3 — anti-competitive agreements;
  • Sections 5 and 6 — combinations.

In digital markets, dominance may arise from factors such as:

• network effects;
• access to large datasets;
• switching costs;
• ecosystem integration;
• economies of scale;
• consumer defaults;
• interoperability control;
• control over distribution;
• technological barriers to entry; and
• accumulated behavioural data.

III. Intelligent Influence as a Source of Market Power

Traditional market power is often associated with price.

Digital influence infrastructures require a broader analysis.

A platform may exercise power through:

A. Visibility

Who appears first?

B. Salience

Which option receives the most visual prominence?

C. Defaults

Which service is automatically selected?

D. Recommendations

Which products does the system suggest?

E. Personalisation

What does each consumer see?

F. Access

Which competitors receive API or interoperability access?

G. Data

Who obtains behavioural information necessary to improve competing products?

H. Switching friction

How difficult is it for users or sellers to move elsewhere?

Thus:

Control over consumer attention can become a form of competitive power.

IV. Self-Preferencing

One of the most important applications is self-preferencing.

Suppose a dominant platform operates:

• a search engine;
• a marketplace;
• an app store; or
• a recommendation system,

while simultaneously offering its own competing service.

The platform may allegedly manipulate its ranking infrastructure so that its own service receives preferential treatment.

The competitive concern is not simply that the platform owns a competing product.

The concern is:

Whether control over the infrastructure used to allocate visibility is being used to disadvantage competing suppliers.

V. Case Law

1. Google Search (Shopping) — European Commission, Case AT.39740

This is one of the central authorities concerning intelligent influence infrastructure.

The European Commission found that Google had abused its dominance in general search by giving preferential positioning to its own comparison-shopping service while demoting competing comparison-shopping services through its generic search algorithms.

Competition principle

A dominant search infrastructure cannot necessarily use its ranking architecture to give its own downstream service preferential visibility while rivals are subjected to disadvantageous ranking mechanisms.

Importance

The case demonstrates that:

Algorithmic ranking + dominance + self-preferencing + foreclosure

can constitute a competition-law problem.

It also illustrates that visibility itself can be competitively significant.

2. Google Android — European Commission, Case AT.40099

The Google Android decision concerned Google's contractual and technological control over the Android ecosystem.

The Commission examined, among other things:

• pre-installation of Google Search;
• default search arrangements;
• Chrome pre-installation;
• anti-fragmentation restrictions; and
• exclusivity incentives.

The Commission concluded that these arrangements contributed to protecting Google's position in general search and restricted opportunities for competing search services.

Competition principle

An intelligent ecosystem may influence competition through defaults and distribution architecture, even without directly preventing users from downloading competing products.

Relevance to IIIs

This demonstrates the importance of:

Default architecture → consumer inertia → reduced rival visibility → reinforcement of market power.

3. Amazon Marketplace — European Commission

The European Commission's Amazon investigation concerned the relationship between Amazon's marketplace, its retail operations and independent sellers.

Competition concerns included:

• Amazon's use of non-public marketplace seller data;
• Buy Box selection;
• Prime eligibility;
• logistics services; and
• potential preferential treatment of Amazon's own activities.

The Commission ultimately accepted commitments addressing these concerns, including commitments concerning non-public seller data and a more neutral Buy Box process.

Competition principle

A platform that simultaneously:

1. operates the marketplace,
2. collects information from marketplace participants, and
3. competes with those participants,

may possess a structural information advantage.

The Buy Box is particularly significant because algorithmic selection can determine which seller receives the consumer's attention and transaction.

4. FTC v. Amazon

The U.S. Federal Trade Commission and state attorneys general sued Amazon alleging that Amazon maintained monopoly power through a collection of exclusionary practices.

Among the allegations were claims involving:

• search-result manipulation;
• anti-discounting mechanisms;
• marketplace restrictions;
• Prime-related conduct;
• advertising;
• seller dependence; and
• alleged preferential treatment of Amazon's products.

The FTC specifically alleged that Amazon used algorithms to bury sellers offering lower prices and that Amazon's search system could preference Amazon products. These remain allegations in litigation rather than established findings of liability.

Competition principle

This case illustrates how an algorithm can become part of a broader monopolization strategy.

The relevant question is not simply:

"Does Amazon use an algorithm?"

Rather:

"Is algorithmic control being used as part of conduct that excludes competitors or protects monopoly power?"

5. FTC v. Facebook/Meta

The FTC's Facebook litigation concerns alleged maintenance of monopoly power in personal social networking.

The FTC alleged conduct involving:

• acquisitions of Instagram and WhatsApp;
• restrictions on developers;
• API access; and
• interoperability conditions.

The FTC alleges that restrictions on API access could hinder competing applications and reinforce Facebook's position. The litigation has continued through multiple stages, and the allegations should be distinguished from adjudicated findings.

Competition principle

Interoperability can itself be an influence infrastructure.

A dominant platform can influence competitive opportunities by determining:

• who receives access;
• what technical capabilities are available;
• what data can be exchanged;
• which applications can interoperate; and
• on what terms.

6. Google Search Self-Preferencing under the Digital Markets Act

The EU's Digital Markets Act provides a particularly important development because it addresses certain gatekeeper practices through ex ante obligations, rather than relying exclusively on traditional abuse-of-dominance litigation.

In July 2026, the European Commission announced a €460 million fine concerning Google's treatment of its own services in Google Search, finding that Google gave preferential treatment to its own services, including shopping, hotels, transport and sports results.

Competition principle

The DMA specifically addresses the problem of a gatekeeper controlling the infrastructure through which consumers discover competing services.

This represents a movement from:

ex post competition enforcement

towards:

ex ante regulation of digital gatekeeping infrastructure.

7. Google–AI Interoperability and Search Data — 2026

A particularly important recent development concerns AI influence infrastructures.

In July 2026, the European Commission issued binding specification measures concerning Google's Android ecosystem and search-data access. The measures sought to facilitate equal access for competing AI assistants to relevant Android functionality and to provide third-party search engines access to search data that Google Search can collect at scale.

Competition significance

AI assistants increasingly function as:

• search intermediaries;
• recommendation engines;
• shopping agents;
• information filters; and
• decision-making interfaces.

Consequently, control over the AI interface through which consumers reach markets may become an important source of competitive advantage.

VI. Intelligent Influence and Network Effects

Intelligent influence systems can generate positive feedback loops.

For example:

More users

↓

More behavioural data

↓

Better predictions

↓

More relevant recommendations

↓

Higher consumer engagement

↓

More sellers and advertisers

↓

More transactions

↓

More data

This produces a data-network-effect cycle.

The result can be substantial barriers to entry.

A new competitor may possess a technically good algorithm but still struggle because it lacks:

• historical data;
• users;
• advertisers;
• sellers;
• behavioural feedback;
• distribution;
• interoperability; and
• consumer trust.

VII. Algorithmic Personalisation and Discrimination

Intelligent influence infrastructure can also produce discriminatory competitive outcomes.

Examples include:

• showing different prices to different users;
• ranking sellers differently;
• preferential recommendations;
• differential advertising exposure;
• discriminatory access to APIs;
• differentiated search visibility;
• personalised offers; and
• differential commission structures.

Competition authorities must distinguish legitimate personalisation from anticompetitive discrimination.

Personalisation becomes particularly significant where:

the platform controls the market interface and also competes within the market.

VIII. Dark Patterns and Competition

Intelligent influence can operate through choice architecture.

Examples include:

• preselected subscriptions;
• difficult cancellation;
• default payment methods;
• repeated prompts;
• preferential placement;
• confusing comparisons;
• artificial urgency;
• personalised prompts.

The UK Competition and Markets Authority has identified algorithmic choice architecture and ranking systems as potentially relevant to competition, including the use of algorithms to favour a platform's own products.

From a competition-law perspective, the issue is whether such mechanisms:

1. increase switching costs;
2. weaken rivals;
3. reduce consumer choice;
4. reinforce network effects; or
5. prevent effective competition.

IX. Algorithmic Advertising as Influence Infrastructure

Digital advertising is itself an intelligent influence infrastructure.

A typical system operates as:

Advertiser → bidding algorithm → ad exchange → ranking algorithm → consumer interface → behavioural response

The infrastructure can determine:

• who obtains visibility;
• what price advertisers pay;
• which advertisements are shown;
• what consumers see;
• how publishers monetise traffic.

Google's ad-tech activities have therefore attracted competition scrutiny, including an ongoing UK CMA investigation concerning alleged abuse of dominance in ad tech.

The competition problem can become particularly complex when one company controls multiple layers of the advertising supply chain.

X. AI and Intelligent Influence

AI intensifies these competition issues because AI systems can increasingly determine what information consumers receive before consumers even conduct a conventional search.

An AI assistant can:

• select products;
• compare prices;
• recommend suppliers;
• negotiate transactions;
• filter information;
• rank alternatives;
• determine which websites users visit;
• execute purchases; and
• learn from transaction data.

This creates a new potential bottleneck:

The AI interface may become the gateway between consumers and markets.

Consequently, competition law may need to examine whether dominant AI infrastructure can:

• favour its own services;
• suppress rival AI models;
• restrict interoperability;
• deny access to essential data;
• tie AI services to operating systems;
• impose discriminatory access conditions; or
• use accumulated data to foreclose competitors.

The FTC's study of major AI partnerships has similarly examined whether relationships between large cloud providers and AI developers can affect access to computing resources, switching costs and sensitive technical/business information.

XI. Intelligent Influence and Tying

Tying may arise when a dominant undertaking conditions access to one important infrastructure on acceptance of another service.

Examples might include:

Operating system → AI assistant

App store → payment service

Marketplace → logistics

Search engine → specialised search service

Cloud infrastructure → AI model

The Google Android decision demonstrates how contractual conditions surrounding access to the Play Store, Search and Chrome were examined as part of Google's strategy concerning mobile search competition.

XII. Intelligent Influence and Essential Facilities

A digital platform can sometimes become an important access point.

Potential examples include:

• app stores;
• payment rails;
• operating systems;
• cloud computing;
• digital advertising exchanges;
• marketplace infrastructure;
• API gateways;
• search engines;
• AI model interfaces.

Competition law may ask:

Can rivals realistically compete without access to the infrastructure?

Where access is indispensable and legal conditions for an essential-facilities theory are satisfied, refusal or discriminatory access can become competition-law relevant.

However, mere importance is not automatically equivalent to an essential facility. The precise legal test varies by jurisdiction.

XIII. Data as an Influence Asset

Data can strengthen intelligent influence infrastructure in several ways.

1. Training advantage

More data can improve AI models.

2. Prediction advantage

More behavioural data can improve recommendations.

3. Targeting advantage

More information can improve advertising.

4. Personalisation advantage

More data can make switching less attractive.

5. Feedback advantage

More transactions generate additional data.

Thus:

Data can function simultaneously as an input, competitive advantage and barrier to entry.

The Amazon proceedings demonstrate how access to marketplace-generated seller information can become a competition concern when the platform also competes with those sellers.

XIV. Algorithmic Coordination

Intelligent infrastructures can also facilitate coordination among competitors.

Algorithms may:

• monitor rivals continuously;
• detect price changes;
• respond automatically;
• implement common pricing strategies;
• predict competitor behaviour.

This creates two distinct legal situations.

Situation 1 — Explicit coordination

Competitors agree to use an algorithm to implement an anticompetitive agreement.

This can fall within traditional cartel rules.

Situation 2 — Autonomous parallel behaviour

Algorithms independently reach similar pricing outcomes.

This is substantially more difficult.

Competition law must determine whether there is sufficient evidence of:

• communication;
• concerted practice;
• facilitating conduct;
• conscious parallelism; or
• unilateral lawful conduct.

The existence of similar algorithmic outcomes alone does not automatically establish an unlawful cartel.

XV. Market Definition in Intelligent Influence Markets

Traditional market definition becomes difficult.

Consider an AI assistant.

Is the relevant market:

• AI assistants?
• search engines?
• general information services?
• digital recommendation services?
• shopping intermediaries?
• operating systems?

Similarly, a marketplace may simultaneously operate in:

• retail;
• marketplace intermediation;
• logistics;
• advertising;
• payments;
• data services.

Competition analysis therefore increasingly requires ecosystem-based market analysis, while still identifying the legally relevant markets for the particular conduct.

XVI. Multi-Sided Markets

Intelligent influence platforms frequently connect several groups:

Consumers

↕

Platform

↕

Sellers

↕

Advertisers

↕

Developers

The platform may subsidise one side while monetising another.

For example:

Free consumer access → huge user base → valuable behavioural data → advertising revenue.

Therefore, absence of a monetary price to consumers does not necessarily mean absence of market power.

XVII. Consumer Welfare and Innovation

Competition law should distinguish between:

Legitimate intelligent innovation

• better recommendations;
• improved search;
• fraud detection;
• personalised interfaces;
• efficient matching;
• lower transaction costs.

and:

Potentially exclusionary intelligent influence

• systematic self-preferencing;
• discriminatory access;
• exclusionary defaults;
• strategic interoperability restrictions;
• manipulation of rankings to disadvantage rivals;
• exploitative use of competitively sensitive data.

The existence of an algorithm or AI system does not itself establish anticompetitive conduct.

The legal inquiry concerns its competitive effects and the surrounding conduct.

XVIII. Remedies

Competition authorities may employ several remedies.

1. Ranking neutrality

Require objective and non-discriminatory ranking criteria.

2. Data separation

Prevent a platform's retail or downstream division from using sensitive competitor data.

3. Interoperability

Require technical access for competing services.

4. Data portability

Permit users or businesses to transfer relevant data.

5. Choice screens

Reduce the effect of defaults.

6. Non-discrimination

Prevent discriminatory access or ranking.

7. Structural separation

Separate platform infrastructure from competing downstream activities where behavioural remedies are insufficient.

8. Transparency

Require disclosure concerning material ranking or recommendation practices, subject to protection of legitimate trade secrets.

9. Monitoring trustees

Independent monitoring can verify compliance with behavioural commitments. Amazon's EU commitments, for example, included monitoring mechanisms and complaint procedures.

XIX. Emerging Regulatory Model

The regulation of intelligent influence is moving toward three complementary models.

Model 1 — Traditional antitrust

Is there dominance?

↓

Is there exclusionary or exploitative conduct?

↓

What are the competitive effects?

Model 2 — Digital market regulation

Is the undertaking a designated gatekeeper?

↓

Does the conduct violate a specific obligation?

↓

What compliance remedy is required?

Model 3 — Technology-specific governance

AI and digital regulators may additionally examine:

• data access;
• interoperability;
• algorithmic transparency;
• privacy;
• cybersecurity;
• consumer protection;
• AI governance.

This produces a multi-regulator competition environment.

XX. Key Competition Concerns — Summary Table

XXI. Six Core Legal Principles from the Case Law

The cases collectively demonstrate the following principles:

Principle 1 — Ranking can be a competitive resource

Google Shopping demonstrates the importance of algorithmic visibility.

Principle 2 — Defaults can reinforce dominance

Google Android demonstrates the competitive significance of default and pre-installation arrangements.

Principle 3 — Marketplace data can create conflicts of interest

Amazon Marketplace illustrates concerns where a platform simultaneously operates infrastructure and competes with its users.

Principle 4 — API access can determine competitive opportunities

FTC v. Facebook/Meta demonstrates the relevance of interoperability and API restrictions.

Principle 5 — Algorithmic infrastructure can form part of monopolization

FTC v. Amazon illustrates how alleged algorithmic conduct can be considered together with other exclusionary practices rather than in isolation.

Principle 6 — Digital gatekeeper regulation increasingly addresses influence directly

The DMA's Google Search enforcement demonstrates an ex-ante approach to self-preferencing and ranking discrimination.

XXII. Indian Competition-Law Perspective

Under Section 4 of the Competition Act 2002, intelligent influence infrastructures may potentially raise questions concerning:

• unfair or discriminatory conditions;
• unfair or discriminatory prices;
• limitation of production or technical development;
• denial of market access;
• tying;
• leveraging dominance from one market into another.

For digital platforms, the Competition Commission of India may have to examine:

1. relevant market;
2. dominance;
3. network effects;
4. data advantages;
5. entry barriers;
6. consumer dependence;
7. interoperability;
8. ranking and recommendation systems;
9. self-preferencing;
10. foreclosure of competitors.

The same conduct may also implicate the newer regulatory framework concerning digital markets, data protection and consumer protection.

XXIII. Future Competition-Law Questions

Intelligent influence infrastructures will create increasingly difficult questions.

1. AI gatekeepers

If an AI assistant recommends only certain suppliers, who controls market access?

2. Autonomous purchasing

If AI agents purchase products automatically, who determines the ranking criteria?

3. Foundation-model leverage

Can a dominant AI model favour its own downstream services?

4. Synthetic competition

Can AI-generated competitors realistically discipline incumbent platforms?

5. Data asymmetry

Should dominant platforms be required to provide competitors with particular categories of data?

6. Algorithmic neutrality

Can competition law require neutrality without destroying legitimate innovation?

7. Personalised competition

If every consumer receives a different ranking, how should foreclosure be measured?

Conclusion

Intelligent Influence Infrastructures represent a new dimension of market power.

Traditional competition law largely focused on control over:

price + output + physical distribution.

Digital competition increasingly requires attention to:

data + visibility + ranking + recommendation + defaults + interoperability + attention + AI-mediated decisions.

The critical legal distinction is between competition-enhancing intelligence and strategic use of intelligent infrastructure to exclude or disadvantage rivals.

The major authorities—Google Shopping, Google Android, Amazon Marketplace, FTC v. Amazon, FTC v. Facebook/Meta, and the EU's Digital Markets Act enforcement concerning Google Search—show the evolution from conventional dominance analysis toward scrutiny of the infrastructure through which digital markets themselves are organised.

The emerging principle can therefore be expressed as:

Control over the mechanism that determines what consumers see, select and access can become a source of market power comparable in importance to control over price or physical distribution.

For competition law, the central challenge is to ensure that intelligence improves market matching and innovation without becoming an instrument for foreclosure, discrimination, exclusion or durable entrenchment of market power.