Competition Law and Long-Term Evolution of Competition Institutions in AI Civilizations

1. Introduction

Competition law and the long-term evolution of competition institutions in AI civilizations concerns how antitrust authorities, courts, regulators, and economic institutions may need to evolve as artificial intelligence becomes a major component of production, distribution, innovation, and economic decision-making.

An AI-intensive economy may involve:

AI foundation models;

autonomous agents;

robotics;

algorithmic marketplaces;

cloud computing;

specialized AI chips;

automated pricing;

AI-powered financial services;

autonomous supply chains;

AI-generated products;

digital platforms;

machine-to-machine transactions.

Traditional competition institutions remain important, but they may require new technical capabilities and investigative methods.

The long-term objective is to preserve contestable markets, innovation, consumer choice, technological diversity, and fair access to critical infrastructure.

2. Meaning of AI Civilizations

For competition-law purposes, an AI civilization can be understood as a future or highly AI-dependent economic system in which artificial intelligence substantially influences:

production;

consumption;

investment;

logistics;

pricing;

research;

commerce;

resource allocation;

business decisions.

AI may therefore become not merely a product but a fundamental economic infrastructure.

This creates a major institutional question:

Can competition authorities designed for twentieth-century industrial markets effectively supervise twenty-first-century AI ecosystems?

3. Meaning of Competition Institutions

Competition institutions include:

Competition authorities

Courts

Economic regulators

Merger-review bodies

Sector regulators

Consumer-protection agencies

Standard-setting institutions

International cooperation networks

Investigative and technical agencies

Their functions include:

detecting cartels;

investigating abuse of dominance;

reviewing mergers;

defining markets;

imposing remedies;

protecting competitive entry;

monitoring market concentration.

In AI economies, these institutions may need substantially greater technical, economic, and computational capabilities.

4. Why AI Changes Competition Governance

AI can alter competition through several mechanisms.

Traditional economy

Capital → Labour → Production → Distribution → Consumer

AI-intensive economy

Data + Computing + Algorithms + Models + Infrastructure → Automated Production and Distribution

This creates new sources of market power.

A company may gain an advantage from controlling:

training data;

computing resources;

foundation models;

AI talent;

cloud infrastructure;

distribution channels;

user interfaces.

5. Evolution of Competition Institutions

Competition institutions may evolve through several stages.

Stage 1: Traditional Antitrust

Focus on:

cartels;

monopolization;

mergers;

exclusive dealing;

predatory pricing.

Stage 2: Digital Competition

Additional focus on:

platforms;

network effects;

data;

digital advertising;

app stores.

Stage 3: AI Competition Governance

Greater focus on:

foundation models;

compute;

AI agents;

algorithmic coordination;

AI acquisitions;

model distribution.

Stage 4: Autonomous Economic Governance

Potential focus on:

machine-to-machine transactions;

autonomous pricing;

autonomous purchasing;

decentralized AI;

AI-controlled infrastructure.

6. AI and Market Definition

Market definition remains a fundamental competition-law task.

However, AI products can complicate the analysis.

An AI company may simultaneously provide:

search;

advertising;

software;

cloud services;

productivity tools;

AI assistants.

The relevant competitive constraint may therefore come from several adjacent technologies.

Authorities must consider:

functionality;

substitutability;

consumer behaviour;

technological alternatives;

geographic scope;

multi-sided relationships.

7. Computing Power as a Competitive Resource

Advanced AI models may require significant:

GPUs;

data centres;

electricity;

networking infrastructure;

cloud resources.

If access to these resources becomes concentrated, barriers to entry may increase.

Competition institutions may therefore monitor:

exclusive compute arrangements;

discriminatory cloud access;

long-term capacity agreements;

vertical integration;

acquisition of infrastructure providers.

8. Data as a Source of Market Power

AI systems often benefit from extensive datasets.

A potential feedback loop is:

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

This can produce significant competitive advantages.

However, competition authorities should distinguish:

Legitimate data advantage

from

Data-based exclusion.

Possession of large datasets does not automatically constitute an abuse of dominance.

9. AI Foundation Models

Foundation models may become important inputs for many downstream businesses.

A single model provider could potentially serve:

developers;

enterprises;

consumers;

government services;

autonomous agents.

This creates possible competition questions concerning:

licensing;

access;

pricing;

interoperability;

exclusive arrangements;

vertical integration.

10. AI Platforms and Gatekeeping

A company controlling a major AI assistant may become an important gateway between users and businesses.

For example:

Consumer → AI Assistant → Search → Marketplace → Seller → Payment

If the AI system determines which products, services, or suppliers consumers see, its ranking decisions may have significant competitive effects.

Potential concerns include:

self-preferencing;

discriminatory recommendations;

paid ranking;

exclusive partnerships;

restricted access.

11. Algorithmic Collusion

AI systems can independently monitor competitors and change prices.

This creates a difficult competition-law question.

Independent adaptation

Two AI systems independently change prices based upon market conditions.

This is not automatically unlawful coordination.

Coordinated conduct

Businesses deliberately design systems to implement an agreement or exchange competitively sensitive information.

This may raise traditional cartel concerns.

Competition institutions will therefore need sophisticated methods for distinguishing algorithmic parallelism from unlawful coordination.

12. Autonomous Agents

AI agents may eventually:

purchase goods;

negotiate prices;

select suppliers;

manage inventories;

switch providers.

Competition institutions may need to determine who is legally and economically responsible when autonomous agents make commercially significant decisions.

Potentially relevant actors include:

the business deploying the agent;

the developer;

the platform operator;

the data provider;

the infrastructure provider.

Technological autonomy should not automatically eliminate legal accountability.

13. AI Mergers

Merger control may become one of the most important institutional functions.

An established technology company may acquire an AI startup with:

low revenue;

few employees;

limited market share;

but significant:

technology;

data;

patents;

researchers;

user networks;

future competitive potential.

Competition institutions may therefore need to analyze future competition, not only present market share.

14. Killer Acquisitions in AI

AI markets may produce acquisitions where a large firm purchases a potentially disruptive startup before it becomes a significant competitor.

Potentially relevant questions include:

Would the startup likely have become a competitor?

Does the acquisition eliminate an alternative technology?

Does the acquisition increase access to strategically important data?

Does it strengthen an existing AI ecosystem?

Does it increase barriers to entry?

An acquisition is not automatically anticompetitive simply because the target is technologically promising.

15. Interoperability

AI civilization may contain numerous systems that need to interact.

Examples include:

AI assistants;

cloud services;

robotics;

autonomous vehicles;

payment systems;

smart infrastructure.

Competition institutions may consider whether interoperability is necessary to prevent artificial lock-in.

Possible mechanisms include:

APIs;

data portability;

technical standards;

compatibility requirements.

These measures must also account for:

privacy;

cybersecurity;

intellectual property;

safety.

16. Case Law

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

Microsoft is a foundational technology competition case.

The case concerned Microsoft's conduct surrounding the Windows operating system and competing browser technology.

Institutional importance

It demonstrates the importance of competition institutions understanding:

network effects;

platform economics;

technological integration;

exclusionary conduct.

AI relevance

An AI platform may similarly become an infrastructure layer from which adjacent competitors depend.

Case 2: United Brands Company v. Commission, Case 27/76 (1978)

United Brands remains a foundational European competition decision concerning:

market definition;

dominance;

competitive constraints.

AI relevance

AI markets may be difficult to define because a single technology can perform multiple functions.

The case illustrates why authorities must carefully identify the actual competitive market rather than relying solely on product labels.

Case 3: Hoffmann-La Roche & Co. AG v. Commission, Case 85/76 (1979)

The case concerned exclusive arrangements involving a dominant undertaking.

The Court emphasized the special responsibility of dominant firms not to impair genuine competition.

AI relevance

Similar concerns could arise from:

exclusive AI distribution;

exclusive cloud agreements;

preferential AI integration;

contractual restrictions on developers.

Case 4: AKZO Chemie BV v. Commission, Case C-62/86 (1991)

AKZO established important principles concerning predatory pricing.

AI relevance

AI services may be offered:

free of charge;

below apparent cost;

through heavily subsidized subscriptions.

Competition institutions must determine whether pricing reflects legitimate investment and competition or an exclusionary strategy.

Case 5: Bronner v. Mediaprint, Case C-7/97 (1998)

Bronner addressed refusal of access to a distribution infrastructure.

AI relevance

Future AI markets may contain infrastructure that businesses consider indispensable, such as:

cloud computing;

model-access infrastructure;

AI distribution platforms.

The case illustrates the exceptional nature of compelled access under traditional competition principles.

Case 6: Intel Corp. v. Commission, Case C-413/14 P (2017)

Intel concerned rebates by a dominant undertaking.

Institutional importance

The case illustrates why competition authorities need sophisticated economic analysis when examining conduct capable of excluding competitors.

AI relevance

Similar questions could arise with:

cloud rebates;

AI-platform discounts;

developer incentives;

volume discounts;

ecosystem loyalty programs.

Case 7: Google and Alphabet v. Commission, Case T-612/17 (Google Shopping) (2021)

Google Shopping concerned treatment of competing comparison-shopping services in search results.

AI relevance

AI assistants may become the next major interface for consumers.

An AI system that controls recommendations could potentially influence which businesses receive visibility.

This makes the principles surrounding:

ranking;

self-preferencing;

platform power;

access to consumers

highly relevant to future AI markets.

Case 8: Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985)

Aspen Skiing concerned refusal-to-deal conduct.

AI relevance

Comparable questions may arise where a dominant AI platform terminates access to:

APIs;

model interfaces;

marketplaces;

technical infrastructure.

The case does not establish that every refusal to deal is unlawful.

17. Development of AI Competition Authorities

Future competition institutions may require specialized departments.

A. AI Economics Unit

Experts in:

industrial organization;

econometrics;

platform economics;

dynamic competition.

B. Technical AI Unit

Experts in:

machine learning;

algorithms;

cloud computing;

cybersecurity.

C. Digital Evidence Unit

Experts capable of examining:

logs;

source code;

model behaviour;

internal communications;

data flows.

D. Merger Intelligence Unit

Focused on:

startup acquisitions;

emerging technologies;

potential competition.

18. AI-Assisted Competition Enforcement

Competition authorities themselves may use AI.

AI could assist in:

identifying suspicious pricing patterns;

detecting potential bid-rigging;

analyzing millions of documents;

identifying unusual merger patterns;

monitoring market concentration;

mapping corporate relationships.

However, human oversight remains important.

AI-generated investigative conclusions should be subject to:

verification;

procedural safeguards;

explainability appropriate to the decision;

judicial review.

19. Competition Courts in AI Economies

Courts may increasingly encounter disputes involving:

algorithmic evidence;

AI-generated records;

complex economic models;

source-code evidence;

automated decision systems.

Judges may therefore require access to:

technical experts;

economic experts;

independent assessments.

The judicial role remains important because competition enforcement ultimately requires legally accountable decision-making.

20. Ex-Ante Competition Institutions

Future competition institutions may combine traditional enforcement with preventive regulation.

Possible tools include:

Interoperability requirements

To reduce ecosystem lock-in.

Data portability

To facilitate switching.

Merger notification

For strategically significant transactions.

Non-discrimination obligations

For critical digital gatekeepers.

Access obligations

Where exceptional legal conditions are satisfied.

Transparency

For important platform practices.

21. Competition and Innovation

AI competition policy must preserve incentives for technological innovation.

A company that develops a superior AI model should be able to benefit from its innovation.

Competition institutions should not treat:

successful innovation = unlawful dominance.

The legal concern arises when market power is maintained or extended through conduct that unlawfully restricts competition.

22. AI Standards

Standard-setting may become increasingly important.

Potential standards may concern:

AI interoperability;

data formats;

cybersecurity;

model interfaces;

autonomous vehicles;

robotics.

Standards can increase competition by enabling compatibility.

However, companies may potentially use standards strategically to exclude alternative technologies.

Competition institutions should therefore monitor standard-setting processes where appropriate.

23. International Institutional Cooperation

AI markets are global.

An AI company may simultaneously operate in:

United States;

European Union;

United Kingdom;

India;

UAE;

Singapore;

Japan.

Competition institutions may therefore cooperate on:

cross-border mergers;

digital-platform investigations;

cartel detection;

AI infrastructure;

international evidence gathering.

Different jurisdictions will nevertheless retain distinct legal standards.

24. UAE Perspective

For the UAE, the evolution of competition institutions is relevant to sectors such as:

artificial intelligence;

fintech;

telecommunications;

e-commerce;

logistics;

cloud computing;

digital payments;

smart-city infrastructure;

autonomous transportation.

The UAE's principal competition framework is associated with Federal Law No. 4 of 2012 on the Regulation of Competition and its implementing framework.

Future institutional questions may include:

How should AI markets be defined?

How should AI-related mergers be reviewed?

How should algorithmic conduct be investigated?

How should dominant AI platforms be monitored?

How should technical evidence be evaluated?

How should competition policy interact with national AI-development objectives?

25. Institutional Independence

Effective competition governance requires institutions capable of making decisions based upon:

evidence;

economic analysis;

competition principles;

applicable law.

Competition institutions should be sufficiently independent to investigate powerful firms while maintaining appropriate accountability through:

judicial review;

procedural safeguards;

transparent decision-making;

statutory mandates.

26. Long-Term Institutional Roadmap

A possible development model is:

Phase 1 — Strengthen traditional antitrust

Improve cartel, dominance and merger enforcement.

Phase 2 — Build digital expertise

Develop platform, data and algorithm expertise.

Phase 3 — Establish AI monitoring

Monitor foundation models, compute, cloud and AI ecosystems.

Phase 4 — Develop autonomous-market expertise

Study machine-to-machine transactions and autonomous agents.

Phase 5 — Continuous competition governance

Use ongoing market monitoring and early-warning mechanisms.

27. Major Institutional Challenges

1. Technical complexity

AI systems can be difficult for traditional legal institutions to understand.

2. Speed of innovation

Technology can develop faster than legislation.

3. Information asymmetry

Companies may possess substantially more technical information than regulators.

4. Cross-border operations

AI markets rarely respect national boundaries.

5. False positives

Aggressive innovation could be incorrectly treated as exclusion.

6. False negatives

Authorities may intervene too late after market power becomes entrenched.

7. Resource constraints

Effective AI enforcement requires highly specialized personnel.

28. Key Principles

Long-term AI competition institutions should follow these principles:

Technology neutrality

Evidence-based enforcement

Protection of dynamic competition

Preservation of market contestability

Proportionate intervention

Careful merger review

Technical competence

Human accountability

International cooperation

Continuous market monitoring

29. Quick Revision Table

30. Conclusion

The long-term evolution of competition institutions in AI civilizations requires competition authorities and courts to evolve alongside the technologies they regulate.

Traditional principles concerning dominance, exclusion, pricing, access, vertical integration, mergers and platform power remain relevant. Cases such as Microsoft, United Brands, Hoffmann-La Roche, AKZO, Bronner, Intel, Google Shopping and Aspen Skiing provide important foundations.

However, AI economies introduce new competitive resources—including computing power, foundation models, training data, algorithms, AI agents and technological interfaces. Competition institutions will consequently need greater technical expertise, better digital evidence capabilities, more sophisticated merger analysis and continuous market monitoring.

The long-term objective should not be to prevent technological concentration merely because it exists. Rather, it should be to ensure that successful innovation does not become an unnecessary and permanent barrier to future innovation, entry and competitive choice.