Competition Law and Long-Term Competition Frameworks for Autonomous Societies

1. Introduction

Long-term competition frameworks for autonomous societies concern the design of competition-law institutions and rules for an economy in which increasingly large parts of commercial activity are performed or assisted by artificial intelligence, autonomous agents, algorithms, robotics, digital platforms, automated marketplaces, smart infrastructure and interconnected economic systems.

An autonomous society does not mean that human beings disappear from economic activity. Rather, many commercial decisions may increasingly be performed automatically, such as:

purchasing;

pricing;

advertising;

investment allocation;

logistics;

inventory management;

supplier selection;

contract matching;

financial transactions;

recommendation;

negotiation.

This creates a long-term competition-law question:

How should competition law preserve market rivalry when the mechanisms determining competition are increasingly automated and interconnected?

The objective is not to prevent automation. The objective is to ensure that automation does not become a mechanism for permanent market concentration, exclusion of competitors or loss of consumer choice.

2. Meaning of a Long-Term Competition Framework

A traditional competition framework may focus primarily on existing conduct.

A long-term framework additionally considers:

future market structure;

technological development;

potential competitors;

artificial intelligence;

data;

network effects;

interoperability;

autonomous decision-making;

ecosystem power;

merger strategies;

infrastructure dependence.

Therefore:

Long-term competition framework = rules + institutions + monitoring + enforcement + future-oriented market analysis designed to preserve competition over an extended period.

3. Why Autonomous Societies Create New Competition Problems

Autonomous economic systems may generate significant efficiency.

For example:

AI purchasing agent → compares thousands of products → negotiates automatically → purchases → arranges delivery.

This can reduce:

search costs;

transaction costs;

information asymmetry;

administrative costs.

However, if one AI agent controls access to most consumers, the same efficiency can create a powerful bottleneck.

The system could determine:

which firms appear;

which products are recommended;

which payment systems are used;

which suppliers receive orders;

which logistics provider receives business.

Thus, efficiency and concentration can arise simultaneously.

4. Main Objectives

A long-term competition framework should seek to preserve:

1. Contestability

New firms should have a realistic opportunity to enter.

2. Consumer choice

Consumers should have meaningful alternatives.

3. Innovation

Competition should encourage technological development.

4. Market access

Businesses should not be unnecessarily excluded from important platforms.

5. Interoperability

Different technological systems should be able to interact where appropriate.

6. Data mobility

Users and businesses should not be unnecessarily locked into one ecosystem.

7. Competitive neutrality

Government-supported and private businesses should compete under appropriate conditions.

8. Dynamic competition

The framework should protect future competition, not only current rivalry.

5. Autonomous Commercial Decision-Making

Autonomous systems may make decisions concerning:

prices;

inventory;

advertising;

procurement;

lending;

product placement;

supplier selection;

consumer recommendations.

This raises the question of accountability.

The fact that a decision is produced by an algorithm does not automatically remove the responsibility of the undertaking that designed, deployed or controlled the system.

Competition law therefore needs to examine:

Who designed the system?

Who supplied the data?

Who controlled its objectives?

Who benefited from the conduct?

Was the conduct foreseeable or deliberately engineered?

6. Market Definition

Traditional market definition may become difficult in autonomous economies.

An AI ecosystem may simultaneously provide:

search;

advertising;

cloud computing;

digital payments;

shopping;

logistics;

AI assistance.

Competition authorities may therefore need to analyse:

Product market

What service is actually being supplied?

Geographic market

Where does competition occur?

Platform relationships

Which groups depend upon one another?

Ecosystem

How do different markets interact?

Potential competition

Could an emerging technology become a substitute?

7. Dynamic Market Definition

Long-term competition analysis should not assume that today's market boundaries will remain unchanged.

For example:

Search → AI assistant → AI shopping → payment → autonomous purchasing

could transform several formerly separate markets.

A competition authority should therefore consider:

technological convergence;

substitution;

innovation;

potential competitors;

ecosystem expansion.

8. Network Effects

Network effects can create self-reinforcing growth.

For example:

More users → more transactions → more data → better AI → more users.

A successful autonomous platform may therefore acquire an advantage that competitors find increasingly difficult to reproduce.

Competition law should distinguish:

legitimate network-based efficiency

from

network effects reinforced through exclusionary conduct.

9. Data as a Competitive Resource

Autonomous societies may generate enormous datasets.

Examples include:

consumer preferences;

purchasing behaviour;

financial information;

search activity;

logistics information;

supplier performance;

pricing information.

Large datasets may improve AI systems and create competitive advantages.

Potential competition concerns include:

exclusive data arrangements;

discriminatory access;

data combination;

data-based foreclosure;

restrictions on portability.

10. Artificial Intelligence and Competition

AI can affect competition at multiple levels.

Input level

chips;

computing power;

cloud infrastructure;

datasets;

specialized talent.

Model level

foundation models;

AI systems;

proprietary algorithms.

Distribution level

search;

applications;

operating systems;

AI assistants.

Commercial level

marketplaces;

advertising;

financial services;

logistics.

Concentration at several levels can create vertical ecosystem power.

11. Autonomous AI Agents

AI agents may eventually act as commercial intermediaries.

Instead of:

Consumer → Seller

the structure could become:

Consumer → AI Agent → Marketplace → Seller.

The AI agent may decide which seller receives the transaction.

This creates a new competitive bottleneck:

Control over the AI interface may mean control over access to consumer demand.

Competition frameworks therefore need to examine whether AI agents:

favour affiliated firms;

exclude competitors;

manipulate rankings;

restrict alternative services;

impose discriminatory conditions.

12. Algorithmic Pricing

Algorithms can automatically alter prices.

For example:

Algorithm A increases price → Algorithm B responds → Algorithm A responds again.

This can produce rapid price adjustments.

Potential issues include:

collusion;

coordination;

predatory pricing;

discriminatory pricing;

exclusionary pricing.

However:

Parallel algorithmic pricing is not automatically proof of unlawful coordination.

Authorities must establish the applicable legal elements.

13. Algorithmic Collusion

Competition law traditionally distinguishes:

Independent conduct

Two firms independently choose similar prices.

from

Coordination

Firms communicate or otherwise participate in prohibited coordination.

Autonomous systems make this distinction more difficult because algorithms may react rapidly to market information.

Long-term frameworks should therefore develop technical capabilities for identifying:

communication;

coordinated strategies;

algorithmic signalling;

suspicious pricing patterns.

14. Self-Preferencing

An autonomous ecosystem may operate:

Marketplace + Own Products + AI Recommendation Engine.

If the AI recommends the ecosystem's own products preferentially, competing suppliers may be disadvantaged.

This issue has become particularly important in digital competition policy.

The EU's Digital Markets Act includes rules addressing preferential treatment of a gatekeeper's own services in covered circumstances.

15. Interoperability

Interoperability allows different systems to interact.

Examples:

AI Agent A ↔ Marketplace B

Payment System A ↔ Platform B

Cloud A ↔ AI System B

Interoperability can reduce:

switching costs;

lock-in;

network barriers;

ecosystem dependence.

But interoperability requirements must also consider:

security;

privacy;

intellectual property;

technical feasibility;

innovation incentives.

16. Data Portability

Data portability permits consumers or businesses to transfer relevant data.

It can facilitate:

switching;

multi-homing;

competition;

entry by new firms.

For autonomous societies, portability becomes especially important because AI personalization may make users increasingly dependent on accumulated data.

17. Multi-Homing

Multi-homing occurs when a consumer or business uses several competing platforms.

For example:

A seller may simultaneously use:

Marketplace A;

Marketplace B;

Marketplace C.

Multi-homing can reduce dependency on one ecosystem.

Competition policy should therefore examine contractual or technical restrictions that unnecessarily prevent multi-homing.

18. Exclusive Dealing

A dominant autonomous ecosystem may require business users to deal exclusively with it.

Potential concerns include:

foreclosure of rivals;

increased entry barriers;

reduced multi-homing;

reduced innovation.

The analysis should consider:

market power;

duration;

market coverage;

efficiencies;

actual or likely foreclosure.

19. Tying and Bundling

An ecosystem could combine:

AI assistant + cloud + payment + marketplace + logistics.

Bundling may generate efficiencies.

However, if a dominant undertaking uses one product to force adoption of another and thereby forecloses competition, competition law may become relevant.

20. Ecosystem Lock-In

Lock-in can arise from:

proprietary data;

technical incompatibility;

high switching costs;

contractual restrictions;

loss of personalization;

loss of transaction history;

incompatible applications.

Long-term competition frameworks should therefore ask:

Can users realistically leave the ecosystem?

21. Merger Control

Merger control becomes particularly important in autonomous economies.

A dominant company might acquire:

AI startups;

robotics companies;

data companies;

cloud businesses;

cybersecurity firms;

emerging payment platforms.

Traditional turnover thresholds may fail to capture some strategically important acquisitions.

Long-term frameworks therefore need to consider:

potential competition;

innovation;

data;

technology;

future market entry.

22. Killer Acquisitions

A killer acquisition occurs where an incumbent acquires an emerging firm in circumstances where the acquisition may eliminate future competitive pressure.

The target may not currently possess a large market share.

Yet it may possess:

promising technology;

patents;

talent;

data;

innovative algorithms.

The long-term question is:

Could the target become a significant competitor if allowed to develop independently?

23. Vertical Integration

An autonomous society may contain companies controlling several layers:

Semiconductors → Cloud → AI → Operating System → Marketplace → Payments.

Vertical integration can produce efficiency.

But it may also create:

discriminatory access;

foreclosure;

self-preferencing;

tying;

raising rivals' costs.

Competition law therefore needs to examine the effects rather than automatically treating vertical integration as unlawful.

24. Important Case Laws

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

Facts

Microsoft had substantial power in the PC operating-system market. The case involved Microsoft's conduct concerning Internet Explorer and competing technologies.

Principle

The court examined exclusionary conduct used to protect Microsoft's position and limit competitive threats.

Relevance

The case illustrates how control over one technological layer can influence competition in adjacent markets.

For autonomous societies:

Core platform → complementary technology → potential competitor

may require careful competition analysis.

25. United Brands Co. v Commission, Case 27/76

Principle

United Brands is a foundational European authority on:

relevant market;

dominance;

market power;

abuse.

Relevance

Even technologically advanced autonomous economies require a basic assessment of whether an undertaking possesses substantial market power.

The case provides an important foundation for understanding dominance before examining more technologically complex conduct.

26. AKZO Chemie BV v Commission, Case C-62/86

Principle

AKZO is an important authority on predatory pricing by a dominant undertaking.

Relevance

Autonomous systems may optimize prices continuously.

Long-term competition analysis must therefore examine whether an automated pricing strategy could:

eliminate competitors;

sustain losses strategically;

exploit cross-subsidies;

foreclose entry.

Automation does not itself determine legality.

27. Bronner v Mediaprint, Case C-7/97

Principle

Bronner concerned refusal of access to a distribution infrastructure and developed important limits on compulsory access under EU competition law.

Relevance

Autonomous societies may contain infrastructure that competitors cannot easily reproduce.

The case demonstrates that:

Dominance does not automatically create an obligation to provide competitors with access to every asset.

The applicable legal conditions must be satisfied.

28. Ohio v. American Express Co., 585 U.S. 529 (2018)

Facts

American Express operated a two-sided payment platform connecting merchants and cardholders.

Principle

The Supreme Court considered the interconnected nature of the two sides of the platform when analysing competitive effects.

Relevance

Autonomous economies can contain multiple interconnected groups:

Consumers ↔ AI Agents ↔ Sellers ↔ Advertisers ↔ Payment Providers.

Competition analysis may therefore need to consider the interaction between different sides of a platform.

29. MOTOE v Elliniko Dimosio, Case C-49/07

Principle

The case involved an organization that exercised regulatory powers while also participating in an economic activity.

The European Court of Justice considered the competition implications of combining regulatory authority with commercial activity.

Relevance

This is useful by analogy for autonomous ecosystems where one organization:

establishes technical rules;

controls access;

sets standards;

and competes with the businesses subject to those rules.

The combination of rule-making and commercial power can create particular competition concerns.

30. Google Shopping Proceedings

The Google Shopping proceedings concerned Google's treatment of its comparison-shopping service in its search ecosystem.

Principle

The proceedings addressed the competitive implications of Google's treatment of its own service compared with competing services.

Relevance

The case is important for understanding self-preferencing in a platform ecosystem.

It illustrates a broader issue:

A platform that controls access to consumers while competing with businesses using that platform may possess special opportunities to influence competition.

31. Google Android — Case AT.40099

Principle

The European Commission's Android proceedings concerned Google's contractual arrangements involving Android devices and related services.

Relevance

The case illustrates how control over one ecosystem layer can influence competition in related markets.

This is relevant to future autonomous societies involving:

operating systems;

AI assistants;

app stores;

search;

cloud;

digital payments.

32. Lessons From the Cases

These cases originate from different legal systems and should be applied according to their respective statutory frameworks.

33. Long-Term Institutional Design

A competition authority for an autonomous society should have five major functions.

1. Market surveillance

Monitor important markets continuously.

2. Technology assessment

Understand emerging technologies before they become dominant.

3. Merger monitoring

Identify acquisitions of potential future competitors.

4. Algorithmic investigation

Develop capacity to analyse automated commercial systems.

5. Remedy monitoring

Ensure that competition remedies remain effective as technology changes.

34. Ex-Ante Regulation

Traditional antitrust is largely:

Conduct → investigation → decision → remedy.

Autonomous economies may require some:

Designation → obligations → monitoring → enforcement.

The Digital Markets Act provides an important example of ex-ante digital-market regulation, particularly for designated gatekeepers and covered core platform services.

This approach is intended to address certain practices before lengthy individual antitrust proceedings are required.

35. Ex-Post Antitrust

Traditional enforcement remains essential for:

cartels;

abuse of dominance;

exclusionary conduct;

anti-competitive agreements;

mergers.

Ex-ante regulation should therefore supplement rather than automatically replace conventional competition law.

36. Competition and Innovation

Innovation is particularly important in autonomous societies.

Competition authorities should examine:

research pipelines;

technological alternatives;

emerging startups;

intellectual property;

AI capabilities;

future substitutes.

A market may appear competitive today while becoming highly concentrated tomorrow.

Therefore:

Innovation competition is a long-term dimension of market rivalry.

37. Competition and Consumer Autonomy

Autonomous systems increasingly make choices for consumers.

Consumers may not personally select:

seller;

price;

product;

payment provider;

logistics provider.

The AI may select them.

Competition law should therefore preserve meaningful consumer ability to:

choose alternatives;

change agents;

switch platforms;

access competing services;

transfer relevant data.

38. Business-User Protection

Businesses can become dependent on autonomous platforms.

For example:

Seller → AI Marketplace → Consumer

The platform may control:

ranking;

advertising;

customer access;

payment;

logistics;

data.

Long-term competition policy should examine whether business users can:

multi-home;

reach customers elsewhere;

obtain relevant data;

use alternative services;

switch platforms.

39. Autonomous Logistics

Autonomous logistics systems may coordinate:

warehouses;

trucks;

drones;

delivery robots;

inventory;

routing.

A single ecosystem could potentially control:

Warehouse + AI Routing + Delivery Network + Marketplace.

This can create efficiency but also potential foreclosure of independent logistics competitors.

40. Autonomous Financial Systems

AI-driven financial ecosystems could control:

payments;

lending;

investment;

insurance;

financial data.

Competition concerns could involve:

tying;

exclusive access;

data concentration;

discriminatory pricing;

platform foreclosure.

Financial regulation and competition law may therefore need to interact.

41. Government Role

Governments may create or support autonomous infrastructure.

Competition neutrality requires attention to:

subsidies;

preferential access;

government procurement;

public data;

state-owned infrastructure.

Government intervention can pursue legitimate public objectives, but long-term competition effects should also be considered.

42. International Cooperation

Autonomous commercial ecosystems may operate across borders.

A single AI platform could have:

developers in one country;

servers in another;

customers worldwide;

payment processing elsewhere.

Competition authorities may therefore require:

information sharing;

coordinated investigations;

merger cooperation;

compatible remedies;

technical cooperation.

43. AI and Competition Authorities

Competition authorities themselves may use AI for:

cartel detection;

merger screening;

price monitoring;

market mapping;

network analysis;

document review.

However, authorities should maintain:

human oversight;

explainability;

procedural fairness;

data security;

auditability.

The enforcement system should not reproduce the opacity it seeks to regulate.

44. Long-Term Competition Monitoring

A useful monitoring system can examine:

Market concentration

Who controls the market?

Entry

Are new competitors appearing?

Switching

Can users leave?

Multi-homing

Can users use several platforms?

Data

Is data concentrated?

Infrastructure

Are essential inputs controlled by a few firms?

Innovation

Are new technologies emerging?

Acquisitions

Are potential competitors being purchased?

Algorithms

Are automated systems producing exclusionary effects?

45. Possible Remedies

Competition authorities may consider, depending on the applicable law:

Behavioural remedies

non-discrimination;

fair access;

contractual restrictions;

transparency.

Technical remedies

interoperability;

APIs;

data portability.

Structural remedies

separation of business units;

divestiture in exceptional cases.

Merger remedies

asset divestiture;

access commitments;

restrictions on certain conduct.

Monitoring

compliance reports;

independent audits;

continuing supervision.

46. Risks of Over-Regulation

Long-term governance must avoid unnecessary intervention.

Potential risks include:

discouraging investment;

reducing innovation;

increasing compliance costs;

protecting inefficient competitors;

creating regulatory uncertainty;

slowing technological development.

Therefore, the existence of a large autonomous ecosystem should not itself establish an antitrust violation.

47. Risks of Under-Regulation

Conversely, delayed intervention can allow:

Network effects + Data + Scale + Lock-in

to produce durable concentration.

Once competitors disappear, restoring competition can become extremely difficult.

Therefore, competition authorities should develop early-warning mechanisms.

48. Long-Term Competition Framework

A comprehensive framework can be represented as:

Step 1 — Identify strategic ecosystems

AI, cloud, payments, logistics, telecommunications, healthcare and other critical sectors.

Step 2 — Map dependencies

Determine which firms depend on which infrastructure.

Step 3 — Measure market power

Examine:

concentration;

entry barriers;

switching costs;

network effects;

data advantages.

Step 4 — Monitor conduct

Look for:

exclusion;

tying;

bundling;

self-preferencing;

exclusivity;

discriminatory access.

Step 5 — Examine future competition

Identify:

startups;

potential substitutes;

emerging technologies;

innovation pipelines.

Step 6 — Review acquisitions

Examine whether transactions eliminate future competitive threats.

Step 7 — Select proportionate remedies

Use the least restrictive effective intervention consistent with the applicable law.

Step 8 — Reassess continuously

Technology and market structures change.

49. Key Principles

The principal principles are:

Automation is not an exemption from competition law.

Large size is not itself unlawful.

Dominance and abuse of dominance are distinct concepts.

Network effects can strengthen market power.

Data can constitute an important competitive resource.

Algorithms can affect competitive conditions.

Self-preferencing may create competition concerns.

Interoperability can reduce lock-in.

Data portability can facilitate switching.

Potential competition can be important in merger analysis.

Vertical integration requires careful effects analysis.

Two-sided markets require consideration of interdependent market participants.

Infrastructure bottlenecks can create special competition concerns.

Ex-ante and ex-post regulation can operate together.

Competition remedies should be proportionate.

Long-term competition policy should preserve innovation and market entry.

50. Conclusion

Long-term competition frameworks for autonomous societies require competition law to evolve from a system focused primarily on individual firms and individual transactions toward one capable of analysing interconnected technological ecosystems.

The major challenges are:

AI + Algorithms + Data + Network Effects + Infrastructure + Automation + Ecosystem Power + Future Competition.

The central objective is not to prevent autonomous commercial systems from developing. Instead, competition policy should ensure that technological progress does not permanently eliminate competitive alternatives.

The most important long-term questions are:

Can new competitors enter?

Can consumers switch?

Can businesses multi-home?

Can data move between systems?

Can competing technologies interoperate?

Can startups grow into independent competitors?

Can dominant ecosystems acquire potential rivals without adequate scrutiny?

Can infrastructure providers compete fairly with businesses dependent upon their infrastructure?

The cases of Microsoft, United Brands, AKZO, Bronner, American Express, MOTOE, Google Shopping and Google Android provide useful comparative foundations for answering these questions.

Quick Revision Formula

Long-Term Competition Framework =

Market Power + Dynamic Competition + Network Effects + Data + AI + Algorithmic Conduct + Interoperability + Portability + Ecosystem Governance + Merger Control + Innovation + Proportionate Remedies

One-Line Exam Definition

Long-term competition frameworks for autonomous societies are enduring legal, economic, technological and institutional mechanisms designed to preserve market contestability, innovation, consumer choice and opportunities for future competitors as commercial decision-making becomes increasingly automated and ecosystem-based.