Competition Law and Intelligent Systems Market Dominance

Introduction

“Intelligent systems” refers broadly to systems using artificial intelligence, machine learning, automated decision-making, predictive analytics, recommendation engines, data-driven optimisation, or algorithmic coordination to provide products or services. Examples include AI search engines, recommendation systems, digital assistants, automated pricing systems, cloud-AI platforms, ad-tech systems, fraud-detection systems, autonomous platforms, and algorithmic marketplaces.

Competition law becomes particularly important where an undertaking controlling an intelligent system possesses substantial market power and uses the system, its data advantages, technical architecture, interoperability rules, or algorithmic decisions to exclude competitors.

The principal competition concerns are:

1. Acquisition or maintenance of dominance through data and network effects.
2. Self-preferencing by an intelligent platform.
3. Tying and bundling AI functionality with another dominant product.
4. Exclusionary interoperability restrictions.
5. Predatory or discriminatory algorithmic pricing.
6. Exclusive access to data, APIs, models, or computing infrastructure.
7. Algorithmic exploitation of network effects and switching costs.
8. Leveraging dominance from one digital market into adjacent markets.
9. Acquisitions of nascent AI competitors.
10. Algorithmic coordination or tacit collusion.

I. Legal Framework

1. Dominance itself is not unlawful

Competition law generally does not prohibit an undertaking merely because it is dominant.

The concern arises when dominance is acquired or maintained through anti-competitive conduct, or when a dominant undertaking uses its market power to foreclose effective competition.

Under the Indian framework, the principal provisions are:

• Competition Act, 2002, Section 4 — abuse of dominant position.
• Section 3 — anti-competitive agreements.
• Section 5 — combinations.
• Section 19 — inquiry into combinations and anti-competitive conduct.
• Section 26 — investigation procedure.
• Section 27 — orders in cases involving abuse of dominance or anti-competitive agreements.

For intelligent systems, Section 4 is particularly significant because the technology can become an instrument through which dominance is exercised.

II. Why Intelligent Systems Can Create Market Dominance

1. Data advantages

An intelligent system may improve as it receives more:

• user queries;
• transaction data;
• behavioural information;
• feedback;
• location information;
• purchasing information;
• performance data; and
• interaction data.

This can produce a data-feedback loop:

More users → more data → better algorithm → better service → more users → more data

A dominant undertaking may therefore obtain an advantage that smaller competitors cannot easily reproduce.

2. Network effects

Many intelligent platforms become more valuable as the number of users increases.

For example:

More users → more interactions → better predictions → more attractive platform → additional users.

Network effects can make entry increasingly difficult.

3. Switching costs

Intelligent systems may become embedded in a user's:

• data history;
• preferences;
• workflow;
• enterprise software;
• cloud environment;
• identity system;
• recommendation profile; and
• API infrastructure.

A competitor may technically exist but nevertheless be unable to attract users because migration is expensive.

4. Algorithmic optimisation

A dominant undertaking may use algorithms to determine:

• rankings;
• search results;
• recommendations;
• prices;
• advertising allocation;
• access conditions;
• visibility;
• commissions; and
• eligibility.

The competition issue is not simply that an algorithm exists. The legal question is how the dominant undertaking uses the algorithm and what competitive effects result.

III. Forms of Abuse

A. Self-Preferencing

A dominant intelligent platform may favour its own:

• products;
• services;
• applications;
• payment systems;
• advertising products;
• AI assistants; or
• affiliated suppliers.

For example, a dominant search platform could theoretically design its ranking algorithm so that its own AI service receives preferential placement.

The relevant questions include:

1. Is the undertaking dominant?
2. Is the platform an important route to customers?
3. Does the algorithm systematically favour affiliated products?
4. Are competing products demoted?
5. Is the conduct capable of excluding equally efficient competitors?
6. Is there an objective justification?

IV. Six Major Case Laws

1. Google Search (Shopping) — European Union

Case: Google Search (Shopping)

The European Commission found that Google had abused its dominant position in general search by giving preferential treatment to its own comparison-shopping service in search-result placement and display while applying less favourable positioning to competing comparison-shopping services.

The case is highly relevant to intelligent systems because ranking algorithms can function as competitive gatekeepers.

Competition-law significance

It demonstrates that:

• dominance in an upstream digital service can affect adjacent markets;
• algorithmic ranking can influence market access;
• preferential treatment can potentially constitute exclusionary conduct;
• visibility is economically important in digital markets.

Intelligent-system application

An AI-powered platform could potentially create similar concerns if its recommendation or ranking algorithm systematically favours affiliated products.

2. Google Android — European Union

Case: Google Android

The European Commission examined Google's position in mobile operating systems and associated restrictions involving:

• Google Search;
• Google Chrome;
• Google Play Store; and
• Android device manufacturers.

The Commission found several practices abusive, including tying-related arrangements and restrictions affecting competing mobile operating systems.

Relevance to intelligent systems

Modern AI assistants increasingly operate through:

• operating systems;
• app stores;
• browsers;
• search engines;
• cloud platforms.

A dominant operating-system provider could potentially leverage control over the operating system to favour its own AI assistant or AI applications.

Principle

Dominance in infrastructure can be leveraged into adjacent technological markets.

3. Microsoft — Internet Explorer

Case: Microsoft Corp. v. Commission

The EU Microsoft proceedings concerned Microsoft's dominant position in PC operating systems and its tying of Internet Explorer with Windows.

The case illustrates the importance of leveraging dominance from one technological layer into another.

Intelligent-system significance

The same conceptual issue can arise where:

Dominant operating system → bundled AI assistant → reduced opportunities for competing AI assistants

or:

Dominant cloud infrastructure → preferential treatment for proprietary AI models.

The competition inquiry would focus on market definition, dominance, foreclosure and competitive effects.

V. 4. United States v. Microsoft Corp.

Case: United States v. Microsoft Corp., 253 F.3d 34 (D.C. Cir. 2001)

Microsoft's conduct concerning Internet Explorer, browser competition, OEM restrictions and the Windows platform was examined under U.S. antitrust law.

The case is particularly significant for understanding platform power and technological leverage.

Relevance to intelligent systems

Intelligent systems are often embedded within larger ecosystems.

A dominant platform could potentially use:

• default settings;
• technical restrictions;
• APIs;
• interoperability;
• contractual restrictions; or
• distribution arrangements

to disadvantage competing AI products.

The Microsoft litigation demonstrates why competition authorities examine not only the immediate product but also control over the surrounding technological ecosystem.

VI. 5. United States v. Google — Search and Advertising

The U.S. antitrust litigation involving Google has addressed alleged exclusionary practices concerning search distribution and related digital markets.

The cases are important for intelligent-system dominance because search is increasingly integrated with:

• generative AI;
• recommendation;
• advertising;
• browser technology;
• mobile operating systems; and
• digital assistants.

Competition-law lesson

Where a company controls a major gateway through which users obtain information, contractual or technical arrangements affecting distribution can have substantial competitive consequences.

For intelligent systems, the equivalent question could be:

Does control over an important AI gateway allow the undertaking to restrict competing AI services from reaching users?

VII. 6. Amazon Marketplace — European Commission

The European Commission's proceedings concerning Amazon's use of marketplace data examined the relationship between Amazon's role as a marketplace operator and its activities as a retailer.

The central concern involved the use of non-public marketplace seller data.

Importance for intelligent systems

This is particularly relevant to AI-driven marketplaces.

A platform may possess information concerning:

• competitor sales;
• consumer demand;
• product performance;
• pricing;
• inventory;
• conversion rates; and
• consumer preferences.

If the platform uses such data to improve its own competing products or services, competition concerns may arise.

Principle

Control over commercially valuable data can strengthen platform power and create conflicts between platform operator and competing users.

VIII. 7. Facebook/Meta — Data and Platform Power

Competition authorities have examined Meta's use of data, platform relationships and digital-market power in multiple proceedings.

The broader competition-law significance is that data access and data accumulation can become competitive advantages, particularly where the platform simultaneously operates as infrastructure and as a downstream competitor.

For intelligent systems, the issue becomes even more important because AI systems can use large datasets to improve:

• prediction;
• personalisation;
• recommendations;
• advertising;
• fraud detection; and
• automated decision-making.

IX. Algorithmic Self-Preferencing

Suppose an AI marketplace has:

• 80% market share;
• millions of users;
• a proprietary recommendation algorithm; and
• its own competing product line.

The algorithm could potentially rank the platform's products above competing products.

The competition analysis would consider:

A. Dominance

Is the platform dominant?

B. Conduct

Does the algorithm systematically prefer affiliated products?

C. Foreclosure

Are competitors deprived of meaningful access to customers?

D. Effects

Does the conduct reduce:

• sales;
• innovation;
• entry;
• consumer choice; or
• quality?

E. Justification

Is there a legitimate technical or consumer-protection reason for the ranking?

X. Intelligent Systems and Refusal to Deal

An intelligent platform may control access to:

• APIs;
• datasets;
• cloud computing;
• model interfaces;
• app stores;
• payment infrastructure;
• identity systems;
• recommendation systems.

If competitors cannot realistically operate without access to the facility, a refusal or discriminatory restriction may raise essential-facility/refusal-to-deal concerns, depending upon the applicable jurisdictional doctrine.

The classic European cases concerning essential facilities include:

Commercial Solvents v Commission

Established important principles concerning refusal to supply by a dominant undertaking.

Bronner v Mediaprint

The Court established a stringent framework for when refusal to provide access to an infrastructure may constitute abuse.

IMS Health v Commission

Concerned access to a protected information structure and the circumstances in which refusal to license can raise competition concerns.

These principles can be adapted to modern digital infrastructure only cautiously; not every important API, dataset or algorithm automatically becomes an essential facility.

XI. Intelligent Systems and Tying

An intelligent-system provider could potentially tie:

AI assistant + operating system

or:

AI model + cloud infrastructure

or:

AI analytics + enterprise software

or:

AI search + advertising platform.

Competition concerns become stronger where:

1. the undertaking is dominant in the tying market;
2. the products are distinguishable;
3. customers are effectively compelled to obtain the tied product;
4. the practice forecloses competing suppliers; and
5. there is no adequate objective justification.

XII. Algorithmic Pricing and Dominance

Intelligent pricing systems create two different competition problems.

1. Unilateral exclusion

A dominant undertaking may use algorithms to:

• discriminate against rivals;
• target competitors;
• selectively reduce prices;
• increase rivals' costs.

2. Algorithmic coordination

Competitors may use pricing algorithms that respond automatically to each other's prices.

This creates a difficult question:

When does algorithmic interdependence become unlawful coordination?

The mere use of similar pricing software does not automatically establish an unlawful agreement. Competition law generally requires careful examination of communication, coordination, concerted practices and the applicable legal standard.

XIII. Intelligent Recommendation Systems

Recommendation algorithms can determine what consumers see.

Examples include:

• search rankings;
• product recommendations;
• video recommendations;
• music recommendations;
• travel rankings;
• news feeds;
• app recommendations.

A dominant platform may therefore exercise algorithmic gatekeeper power.

Possible concerns include:

1. Self-preferencing

Own products receive higher rankings.

2. Demotion

Competitors are algorithmically pushed downward.

3. Data exclusion

Competitors cannot access the data necessary to compete.

4. Interoperability restrictions

Third-party systems cannot connect effectively.

5. Discriminatory access

The platform gives different algorithmic treatment to similarly situated businesses.

XIV. Intelligent Systems and Merger Control

Competition problems do not arise only after dominance has been established.

A major AI company could acquire:

• a foundation-model developer;
• a data provider;
• a chip-design company;
• an AI startup;
• a cloud provider;
• an AI distribution platform.

Authorities may consider whether the acquisition:

• eliminates a potential competitor;
• strengthens data advantages;
• increases barriers to entry;
• forecloses rival AI developers;
• creates vertical leverage;
• increases interoperability risks; or
• reinforces an existing ecosystem.

Therefore, AI concentration can be a merger-control issue as well as an abuse-of-dominance issue.

XV. Barriers to Entry in Intelligent-System Markets

Important barriers include:

XVI. Consumer Harm

Dominance involving intelligent systems can affect consumers through:

• higher prices;
• reduced privacy;
• lower quality;
• reduced innovation;
• fewer choices;
• discriminatory recommendations;
• reduced interoperability;
• weaker service quality;
• reduced transparency; and
• excessive switching costs.

Importantly, a product being free does not necessarily mean competition law is irrelevant.

Competition can occur through:

• quality;
• privacy;
• innovation;
• functionality;
• interoperability; and
• data practices.

XVII. Innovation Competition

AI markets are particularly sensitive to innovation foreclosure.

A dominant incumbent may have incentives to prevent competitors from developing:

• superior AI models;
• alternative recommendation engines;
• new data architectures;
• competing assistants;
• specialised AI applications.

Competition authorities may therefore examine not merely present prices but also:

whether conduct prevents future technological competition.

XVIII. Relevant Market Definition

Traditional market definition can be difficult.

Potential markets include:

Product market

• general search;
• AI search;
• foundation models;
• cloud AI;
• AI assistants;
• recommendation services;
• AI-enabled advertising;
• enterprise AI software.

Geographic market

The relevant geographic market may be:

• national;
• regional;
• EEA-wide;
• global; or otherwise defined according to competitive conditions.

Authorities may also consider multi-sided markets, where a platform simultaneously serves:

• consumers;
• advertisers;
• developers;
• merchants; and
• business users.

XIX. Evidence in Intelligent-System Dominance Cases

Competition authorities may examine:

1. Algorithmic logs.
2. Ranking changes.
3. Internal communications.
4. API-access records.
5. Data-access policies.
6. Model-training information.
7. User switching rates.
8. Market shares.
9. Entry barriers.
10. Pricing data.
11. A/B testing records.
12. Internal strategy documents.
13. Contracts with distributors.
14. Technical interoperability restrictions.
15. Consumer behaviour.

The technical architecture itself can therefore become competition evidence.

XX. Possible Defences and Objective Justifications

A dominant undertaking may argue that an intelligent-system practice is justified because it:

• improves security;
• prevents fraud;
• protects privacy;
• improves algorithmic accuracy;
• reduces latency;
• protects intellectual property;
• prevents cybersecurity attacks;
• improves interoperability;
• reduces costs;
• enhances consumer experience.

Competition authorities must distinguish legitimate product optimisation from exclusionary conduct.

A beneficial algorithm is not unlawful merely because it disadvantages a competitor.

The central issue is generally whether the conduct constitutes unlawful exclusion or another prohibited form of abuse.

XXI. Remedies

Where unlawful dominance is established, possible remedies can include:

Structural remedies

• divestiture;
• separation of business units;
• prohibition of certain acquisitions.

Behavioural remedies

• non-discriminatory API access;
• interoperability obligations;
• prohibition of self-preferencing;
• data-access requirements;
• non-exclusive distribution;
• transparency obligations;
• modification of contractual restrictions.

Algorithmic remedies

Authorities may potentially require:

• independent auditing;
• monitoring;
• non-discriminatory ranking;
• algorithmic compliance procedures;
• preservation of logs;
• technical access mechanisms.

Such remedies must nevertheless be carefully designed because excessive disclosure of algorithms can create:

• cybersecurity risks;
• intellectual-property concerns;
• gaming of ranking systems; and
• opportunities for manipulation.

XXII. Indian Competition-Law Perspective

For India, intelligent-system dominance should primarily be analysed through Sections 3 and 4 of the Competition Act, 2002, together with the combination provisions.

Relevant forms of conduct can include:

• discriminatory conditions;
• denial of market access;
• unfair conditions;
• predatory pricing;
• tying;
• leveraging;
• exclusive arrangements;
• refusal to deal; and
• exclusionary conduct.

The Competition Commission of India has already developed significant jurisprudence concerning digital platforms, search, e-commerce, app ecosystems, online marketplaces and technology-driven markets.

The analytical framework should therefore combine traditional competition principles with characteristics of digital markets such as:

data + network effects + algorithms + multi-sided platforms + switching costs + interoperability.

XXIII. Consolidated Case-Law Table

XXIV. Key Legal Tests

A useful framework for analysing intelligent-system dominance is:

Step 1 — Define the relevant market

↓

Step 2 — Determine market power

↓

Step 3 — Identify the intelligent-system function

Ranking / pricing / recommendation / data / API / AI model / distribution

↓

Step 4 — Identify exclusionary conduct

Self-preferencing / tying / refusal / discrimination / predation / foreclosure

↓

Step 5 — Establish competitive effects

Entry barriers / foreclosure / innovation reduction / consumer harm

↓

Step 6 — Examine objective justification

Efficiency / security / privacy / technical necessity

↓

Step 7 — Assess proportionality

Is the restriction broader than reasonably necessary?

↓

Step 8 — Determine remedy

Behavioural / interoperability / access / structural / monitoring remedy.

Conclusion

Intelligent systems can create or reinforce market dominance because algorithms, data, network effects, computing infrastructure and digital distribution can mutually reinforce one another. Competition law therefore has to examine not only conventional market shares and prices but also data accumulation, algorithmic control, interoperability, switching costs, ecosystem leverage and innovation foreclosure.

The central competition-law distinction is between legitimate technological superiority and exclusionary exploitation of technological or platform power. A dominant intelligent-system provider may innovate, optimise its products and compete aggressively; the legal concern arises where its control over data, algorithms, infrastructure or distribution is used in a manner that unlawfully restricts competitive access or entrenches dominance.

Thus, intelligent-system market dominance should be analysed through a combination of dominance law, digital-platform economics, data governance, interoperability principles, merger control and algorithmic conduct analysis.