Competition Law and Intelligent Systems Market Dominance

Introduction

Intelligent systems are technology systems that use artificial intelligence, machine learning, automated decision-making, predictive analytics, recommendation engines, natural-language processing, computer vision, or adaptive algorithms to perform or influence commercial decisions.

Examples include:

• AI-powered search and recommendation systems;
• intelligent pricing and bidding systems;
• algorithmic advertising platforms;
• AI-based financial and credit-scoring systems;
• autonomous logistics and procurement platforms;
• smart-device ecosystems;
• AI cloud and foundation-model services;
• intelligent marketplaces and matching platforms.

Competition law becomes particularly important where an undertaking controlling an intelligent system acquires market power or dominance and uses technological advantages, data, algorithms, interoperability restrictions, ecosystem control, or network effects to exclude competitors.

The central issue is not simply that an undertaking possesses sophisticated AI. Rather, the question is whether the intelligent system gives the undertaking substantial market power and whether that power is being maintained or extended through anticompetitive conduct.

1. Meaning of Market Dominance in Intelligent Systems

Market dominance generally refers to a position of economic strength that enables an undertaking to behave to an appreciable extent independently of competitors, customers, or consumers.

In intelligent-system markets, dominance may arise from a combination of:

1. Large datasets
2. Superior algorithms
3. Computational infrastructure
4. Network effects
5. User lock-in
6. Interoperability advantages
7. Control over distribution channels
8. Brand and ecosystem effects
9. Access to specialised computing resources
10. Continuous algorithmic learning

An AI system may therefore become difficult to challenge even where competitors can technically develop similar software.

2. Relevant Market

Determining the relevant market is particularly difficult for intelligent systems.

A. Product-market definition

Authorities may examine whether the relevant market consists of:

• general search services;
• online advertising;
• AI assistants;
• recommendation services;
• cloud computing;
• foundation models;
• AI infrastructure;
• AI-enabled enterprise software;
• app distribution;
• digital marketplaces;
• data-related services.

The analysis may distinguish between traditional products and AI-enhanced products.

For example, an AI recommendation service may compete not merely with other recommendation engines but with alternative mechanisms through which consumers discover products.

B. Geographic market

The geographic market may be:

• national;
• regional;
• global; or
• platform-specific.

Digital services can operate globally, but regulatory conditions, language, data restrictions, privacy requirements and consumer behaviour can make markets geographically narrower.

3. Sources of Dominance in Intelligent Systems

3.1 Data Advantage

AI systems often improve as they obtain more:

• consumer data;
• transaction data;
• behavioural information;
• search queries;
• location data;
• purchasing histories;
• interaction data.

A dominant platform may therefore possess a data feedback loop:

More users → more data → better AI → better service → more users → still more data.

This can produce substantial barriers to entry.

However, possession of large datasets alone does not necessarily establish dominance. Competition authorities must consider whether the data is:

• unique;
• commercially valuable;
• difficult to replicate;
• portable;
• substitutable; and
• necessary for effective competition.

4. Algorithmic Network Effects

Intelligent systems frequently benefit from network effects.

For example:

More sellers → more products → more consumers → more transactions → more data → better recommendations → more sellers.

This can create self-reinforcing market power.

Competition concerns arise when the dominant platform deliberately strengthens these network effects by:

• restricting interoperability;
• preventing data portability;
• imposing exclusivity;
• disadvantaging competing services;
• tying complementary products;
• manipulating rankings;
• restricting APIs.

5. Algorithmic Self-Preferencing

A dominant intelligent platform may operate both:

1. an infrastructure/platform service; and
2. competing downstream services.

The platform's algorithm may then systematically favour its own products.

Examples include:

• search results favouring the platform's own service;
• marketplace algorithms promoting affiliated sellers;
• recommendation engines favouring proprietary products;
• app stores prioritising first-party applications.

This can constitute a competition concern where the conduct excludes equally efficient competitors or exploits control over an important platform.

6. Intelligent Systems and Refusal of Access

A dominant undertaking may control:

• an AI API;
• an essential dataset;
• cloud infrastructure;
• computing capacity;
• interoperability protocols;
• an AI marketplace;
• a critical digital interface.

Refusing access, degrading access, or supplying access on discriminatory conditions may raise refusal-to-deal or essential-facility concerns.

The analysis generally requires careful consideration of:

• indispensability;
• feasibility of replication;
• elimination of effective competition;
• objective justification; and
• consumer harm.

7. Algorithmic Pricing and Dominance

Intelligent pricing systems can analyse:

• competitor prices;
• demand;
• inventory;
• consumer behaviour;
• elasticity;
• historical transactions.

A dominant firm could potentially use algorithmic pricing to:

• discriminate between customers;
• implement exclusionary pricing;
• engage in targeted predation;
• increase switching costs;
• coordinate indirectly with competitors.

Importantly, algorithmic conduct does not escape competition law merely because a computer or AI system implements it.

8. Predatory Pricing by Intelligent Systems

Traditional predatory pricing involves deliberately pricing below an appropriate cost benchmark to eliminate competitors and subsequently recover losses.

AI can make predatory strategies more sophisticated.

An intelligent system can identify:

• vulnerable competitors;
• geographical areas;
• customer groups;
• periods of weak demand;
• competitor financial constraints.

Competition authorities may therefore examine whether algorithmically generated prices are part of an exclusionary strategy.

9. Exclusive Dealing and Intelligent Ecosystems

Dominant AI platforms may require customers to use:

• proprietary APIs;
• proprietary cloud infrastructure;
• proprietary payment systems;
• proprietary data formats;
• proprietary software environments.

This can create ecosystem foreclosure.

The concern becomes stronger where switching to competing systems requires substantial:

• technical migration;
• data conversion;
• retraining;
• employee retraining;
• contractual renegotiation.

10. Tying and Bundling

Intelligent systems are often embedded within larger ecosystems.

A dominant undertaking may bundle:

AI assistant + operating system + browser + cloud + search + advertising.

Competition concerns arise if customers effectively cannot obtain one service without taking another.

The principal questions include:

1. Are there two separate products?
2. Does the undertaking possess dominance in the tying product?
3. Is the customer compelled to obtain the tied product?
4. Does the practice foreclose competitors?
5. Is there an objective justification?
6. Are consumers harmed?

11. Interoperability and API Restrictions

Interoperability is particularly important for AI systems.

A dominant undertaking may control an API necessary for competitors to interact with:

• a platform;
• operating system;
• database;
• cloud service;
• smart-device ecosystem.

Restrictions may include:

• denying API access;
• throttling competitors;
• charging discriminatory fees;
• withholding technical information;
• changing technical standards without adequate notice.

Such conduct can increase rivals' costs and protect the incumbent's market position.

12. Consumer Lock-In

Intelligent systems can create significant switching costs.

A consumer may have accumulated:

• personalised recommendations;
• AI conversation history;
• training preferences;
• stored data;
• automated workflows;
• device integrations.

The more personalised the system becomes, the greater the potential switching cost.

This can create behavioural lock-in, even where technically equivalent alternatives exist.

13. Dominance Through Vertical Integration

AI markets can contain several vertically connected levels:

Semiconductor → cloud computing → foundation model → AI application → distribution platform → consumer.

A vertically integrated undertaking may have incentives to disadvantage independent competitors at downstream or upstream levels.

Potential practices include:

• discriminatory cloud pricing;
• preferential access to computing resources;
• bundling;
• exclusive arrangements;
• discriminatory API terms;
• interoperability restrictions;
• preferential treatment of affiliated applications.

14. Competition Law Case Laws

The following cases are particularly useful for understanding dominance in intelligent-system markets. Several arise from digital-platform markets rather than AI specifically, but their legal principles are highly relevant to AI-driven systems.

Case 1: United States v. Microsoft Corp. — 253 F.3d 34 (D.C. Cir. 2001)

Facts

Microsoft possessed substantial power in the market for Intel-compatible PC operating systems.

It used contractual and technological strategies involving Internet Explorer and competing browsers.

Decision

The D.C. Circuit upheld important findings that Microsoft had unlawfully maintained monopoly power through exclusionary conduct.

Relevance to Intelligent Systems

The case demonstrates that dominance can be reinforced through control over an important technological platform.

For intelligent systems, similar concerns can arise where an AI platform controls:

• operating-system access;
• APIs;
• distribution;
• default settings;
• interoperability.

Principle

A dominant technological platform cannot necessarily use its control over an adjacent technological layer to exclude competitors.

Case 2: United States v. Google LLC — Search and Search Advertising

The U.S. Google search litigation concerns Google's conduct in maintaining its position in general search services and search advertising.

Relevance

The case illustrates the importance of:

• defaults;
• distribution agreements;
• scale;
• data advantages;
• search quality;
• network effects.

For intelligent systems, default placement of an AI assistant or search engine can potentially create substantial competitive advantages.

Principle

Competition analysis may examine whether contractual or distribution arrangements reinforce an already-established technological position.

Case 3: Google Shopping — European Commission

The European Commission found that Google had abused a dominant position in general search by systematically giving prominent placement to its own comparison-shopping service while demoting competing comparison-shopping services.

Relevance

This is particularly important for AI recommendation systems.

An intelligent platform could potentially:

rank → recommend → personalise → promote its own service.

If competitors depend upon the platform for consumer discovery, algorithmic self-preferencing may have significant foreclosure effects.

Principle

A dominant platform's control over ranking and visibility can create competition concerns when that control is used to favour its own downstream service.

Case 4: Google Android — European Commission

The European Commission found that Google imposed certain contractual restrictions relating to Android devices, including practices involving search and browser distribution.

Relevance

The case demonstrates how dominance can extend across an ecosystem.

An intelligent-system provider controlling:

• operating systems;
• app distribution;
• search;
• AI assistants;

may possess opportunities to reinforce market power through contractual and technical integration.

Principle

Competition authorities can examine the cumulative effect of contractual restrictions across interconnected digital markets.

Case 5: Google Android Auto / Enel X — European Commission

The European Commission investigated Google's restrictions concerning interoperability with Android Auto and ultimately addressed concerns regarding access for competing applications.

Relevance

This is directly relevant to intelligent-system interoperability.

AI platforms increasingly depend upon integration with:

• vehicles;
• smart homes;
• mobile devices;
• enterprise systems;
• IoT platforms.

Restricting interoperability can prevent competitors from reaching users.

Principle

A dominant digital ecosystem may raise competition concerns when technical restrictions prevent competing services from interoperating with the platform.

Case 6: Apple App Store — European Union Competition Proceedings

The European Commission's proceedings concerning Apple's App Store have examined Apple's control over app distribution and payment mechanisms.

Relevance

AI applications increasingly reach users through:

• app stores;
• mobile operating systems;
• payment platforms;
• device ecosystems.

A dominant platform can potentially use its gatekeeper position to impose conditions on competing AI applications.

Competition Issues

Potential concerns include:

• mandatory payment systems;
• restrictions on alternative distribution;
• discriminatory commercial conditions;
• limitations on communication with users;
• preferential treatment of affiliated services.

Case 7: Intel v. European Commission — Case C-413/14 P

Facts

Intel was accused of using rebates to restrict competition in the market for x86 central processing units.

The Court of Justice emphasised the importance of examining whether rebate practices were capable of producing anticompetitive foreclosure effects.

Relevance to Intelligent Systems

AI infrastructure depends heavily on specialised processors and computing resources.

Similar principles may become relevant where a dominant infrastructure provider uses:

• loyalty rebates;
• exclusive arrangements;
• conditional discounts;
• bundled computing resources;

to foreclose competing AI infrastructure.

Principle

The economic effects and foreclosure capability of exclusionary pricing practices are important in assessing abuse of dominance.

Case 8: Bronner v Mediaprint — Case C-7/97

Facts

The case concerned access to a newspaper home-delivery system controlled by an established operator.

Principle

The Court adopted a demanding standard for treating refusal of access to infrastructure as abusive.

Among other considerations, the facility must be effectively indispensable and duplication must not be realistically feasible.

Relevance

The same reasoning can be important for:

• proprietary AI datasets;
• AI infrastructure;
• cloud computing;
• specialised AI interfaces;
• essential APIs.

Not every commercially useful AI resource constitutes an essential facility.

15. Key Competition Issues in Intelligent Systems

16. The Role of Data Feedback Loops

One of the most important characteristics of intelligent systems is the feedback loop.

A simplified model is:

Users

↓

Interactions

↓

Data

↓

Training

↓

Improved AI

↓

Better recommendations/services

↓

More users

↓

More data

This can create a substantial barrier to entry.

Competition authorities should therefore examine not only today's market share but also whether a dominant undertaking has created a self-reinforcing competitive advantage.

17. AI and Essential Facilities

An intelligent-system resource may become particularly important where competitors cannot realistically reproduce it.

Potential examples include:

• unique datasets;
• specialised computing infrastructure;
• critical APIs;
• interoperability interfaces;
• dominant digital identity systems;
• platform access mechanisms.

However, treating an AI resource as an essential facility requires careful legal analysis.

The fact that an input is expensive or commercially important does not automatically make it indispensable.

18. Intelligent Systems and Merger Control

Market dominance may also arise through acquisitions.

A large technology undertaking could acquire:

• an AI start-up;
• a specialised model developer;
• a data provider;
• an AI infrastructure company;
• an AI distribution platform.

Competition authorities may consider:

Horizontal effects

Would the merger remove a significant competitor?

Vertical effects

Could the merged entity restrict competitors' access to inputs?

Conglomerate effects

Could the undertaking leverage power from one digital market into another?

Innovation effects

Would the transaction eliminate an important source of future innovation?

This is particularly significant because an emerging AI competitor may have limited current revenue despite possessing significant competitive potential.

19. Indian Competition Law Perspective

In India, intelligent-system dominance is principally analysed under the Competition Act, 2002, particularly the framework concerning:

• relevant market;
• dominant position;
• abuse of dominant position;
• exclusionary conduct;
• discriminatory conduct;
• unfair conditions;
• denial of market access;
• tying and leveraging.

The Competition Commission of India (CCI) has increasingly examined digital-platform characteristics such as:

• network effects;
• data;
• switching costs;
• multi-sided markets;
• platform dependence;
• algorithms;
• interoperability.

The analytical challenge is that traditional indicators such as market share may not fully capture the competitive importance of data, ecosystem integration and technological control.

20. Evidentiary Problems

Intelligent systems create distinctive evidentiary challenges.

Competition authorities may need to investigate:

• source code;
• model architecture;
• training datasets;
• model cards;
• algorithmic logs;
• API records;
• ranking methodologies;
• internal communications;
• pricing outputs;
• experimentation records;
• A/B testing;
• technical documentation.

An authority may need to determine whether apparently neutral algorithmic decisions actually produce systematically exclusionary outcomes.

21. Objective Justification and Efficiency

Not every algorithmic preference is anticompetitive.

A dominant undertaking may argue that an algorithmic decision is justified by:

• security;
• privacy;
• fraud prevention;
• technical compatibility;
• cybersecurity;
• quality control;
• consumer protection;
• system integrity.

The competition analysis should therefore distinguish between:

legitimate technical optimisation

and

strategic exclusion of competitors.

Evidence concerning actual effects and technical necessity can become decisive.

22. Consumer Welfare and Innovation

Competition law in intelligent-system markets should consider both:

Short-term effects

• prices;
• quality;
• choice;
• privacy;
• service availability.

Long-term effects

• innovation;
• entry;
• interoperability;
• technological development;
• diversity of AI models;
• availability of alternative platforms.

A practice may appear beneficial to consumers in the short term while potentially reducing competitive alternatives over time.

23. Compliance Framework for Intelligent-System Dominance

An AI company with substantial market power should establish controls concerning:

1. Ranking

Document why proprietary products receive particular rankings.

2. Pricing

Audit algorithmic pricing for exclusionary outcomes.

3. Data

Ensure data access policies do not unjustifiably discriminate against rivals.

4. APIs

Maintain transparent and objectively justified access criteria.

5. Interoperability

Assess whether technical restrictions are objectively necessary.

6. Contracts

Review exclusivity and loyalty arrangements.

7. Bundling

Assess whether AI services are improperly tied to dominant products.

8. Acquisitions

Evaluate potential competition implications of acquiring emerging AI competitors.

9. Algorithmic auditing

Maintain records capable of explaining significant competitive decisions.

10. Human oversight

Ensure competition-sensitive algorithmic decisions are reviewable.

24. Emerging Doctrine: Algorithmic Dominance

The concept of dominance is likely to evolve from simple market-share analysis toward a broader assessment of technological control.

A sophisticated assessment may examine:

Data + Algorithms + Compute + Network Effects + Distribution + Ecosystem + Switching Costs + Interoperability

Together, these factors may create durable market power even where the underlying AI service is nominally free.

25. Important Legal Principles from the Cases

26. Conclusion

Competition law and intelligent systems intersect most strongly where technological capabilities become sources of durable market power.

The principal concerns are not simply the use of AI but the possibility that a powerful undertaking may combine:

data + algorithms + compute + network effects + ecosystem control + interoperability restrictions + contractual restrictions

to protect or extend its market position.

The most important competition-law questions are therefore:

1. What is the relevant intelligent-system market?
2. Does the undertaking possess substantial market power?
3. What technological or economic factors create that power?
4. Is the conduct exclusionary or exploitative?
5. Does it foreclose competing AI systems?
6. Are competitors denied access to essential inputs or interfaces?
7. Does the conduct produce consumer or innovation harm?
8. Is there an objective technical or efficiency justification?

The developing law of digital platforms—particularly Microsoft, Google Shopping, Google Android, Android Auto, Apple App Store, Intel and Bronner—provides the conceptual foundation for analysing future cases involving AI-driven recommendation, pricing, search, infrastructure, data and decision-making systems.