Competition Law and Future Institutional Models for Autonomous Economies

1. Introduction

The emergence of autonomous economies—economic systems in which AI agents, autonomous software, smart contracts, robotic enterprises, algorithmic marketplaces, decentralized platforms and machine-to-machine transactions perform substantial commercial functions—creates a new institutional problem for competition law.

Traditional competition authorities generally regulate human-controlled firms operating through identifiable corporate structures. Autonomous economies may instead involve:

• AI agents negotiating and purchasing from one another;
• autonomous pricing algorithms;
• decentralized autonomous organizations (DAOs);
• blockchain-based marketplaces;
• algorithmic supply chains;
• machine-controlled financial and commodity markets;
• autonomous procurement and distribution systems;
• foundation-model and AI-agent ecosystems;
• interoperable digital identities and wallets;
• self-executing smart contracts; and
• firms whose competitive strategy is partly determined by continuously learning algorithms.

The central question therefore becomes:

What institutional model should competition authorities adopt when economically significant decisions are increasingly made by autonomous computational systems rather than directly by human managers?

Competition law will probably need to evolve from a predominantly ex-post, firm-centred enforcement model toward a combination of continuous monitoring, algorithmic auditing, interoperability supervision, data governance, technological expertise and coordinated regulatory oversight.

The existing digital-competition cases already provide important building blocks for such an institutional transformation. The EU, UK, US and Australian experiences show increasing attention to self-preferencing, platform ecosystems, data advantages, interoperability and structural market power. The ACCC's 2025 Digital Platform Services Inquiry, for example, expressly examined emerging competition issues involving cloud computing and generative AI and recommended continued monitoring of emerging technologies.

2. Meaning of an Autonomous Economy

An autonomous economy can be understood as an economic environment in which substantial market activity is performed, coordinated or optimized by computational agents capable of acting with limited immediate human intervention.

Main characteristics

A. Autonomous decision-making

AI systems may determine:

• prices;
• quantities;
• suppliers;
• customers;
• advertising allocation;
• credit decisions;
• inventory;
• logistics;
• contract terms; and
• market-entry strategies.

B. Machine-to-machine commerce

Instead of:

Human → Company → Human

the economic structure may become:

AI Agent → AI Platform → Autonomous Supplier → Autonomous Payment System

This makes conventional attribution considerably more difficult.

C. Continuous learning

Traditional commercial decisions are often discrete.

Autonomous systems may instead modify their behaviour every second based upon:

• market conditions;
• competitor prices;
• consumer behaviour;
• available data;
• computational forecasts; and
• reinforcement-learning outcomes.

D. Decentralized organisation

A DAO or blockchain-based network may not have a conventional headquarters, board or centralized management.

E. Algorithmic interdependence

Multiple independent algorithms can observe each other's behaviour and adapt without explicit communication.

This creates difficult questions concerning tacit coordination, algorithmic collusion and responsibility.

3. Why Existing Competition Institutions May Become Inadequate

Traditional competition authorities generally rely upon:

1. complaints;
2. investigations;
3. document requests;
4. witness interviews;
5. economic analysis;
6. forensic evidence;
7. hearings;
8. infringement decisions; and
9. remedies.

These mechanisms remain important, but autonomous economies create additional problems.

The institutional response therefore requires continuous competition governance, rather than enforcement only after harm has occurred.

4. Existing Case Law as a Foundation for Future Institutional Models

Case 1: United Brands v Commission

United Brands Company v Commission, Case 27/76

The European Court of Justice developed important principles concerning:

• dominance;
• market definition;
• economic dependence;
• abusive conduct; and
• exclusionary behaviour.

Relevance to autonomous economies

The case demonstrates that competition law must examine economic power, rather than merely formal corporate structures.

An autonomous economy may therefore require authorities to ask:

Who controls the economically indispensable computational infrastructure?

The relevant power may arise from:

• data;
• computing capacity;
• model access;
• cloud infrastructure;
• AI-agent networks;
• technical standards; or
• interoperability control.

Future authorities could therefore supplement conventional market-share analysis with computational dependency analysis.

5. Case 2: Google Shopping

Google Search (Shopping), European Commission, 2017

The European Commission found that Google had favoured its comparison-shopping service in its general search results. The case became a major reference point for self-preferencing in digital ecosystems.

Importance

Autonomous economies could make self-preferencing substantially more sophisticated.

An AI marketplace might simultaneously control:

• the marketplace;
• ranking;
• recommendation;
• payment;
• advertising;
• logistics; and
• autonomous purchasing agents.

The competition authority may therefore need to examine whether the system's algorithm systematically advantages affiliated services.

Institutional implication

Future competition authorities may require algorithmic neutrality supervision, including:

• ranking audits;
• recommendation audits;
• testing for discriminatory treatment;
• access to relevant technical documentation; and
• continuous monitoring of material algorithmic changes.

The lesson is that competition enforcement increasingly requires technological capabilities beyond conventional legal investigation.

6. Case 3: Google Android

Google Android, European Commission, Case AT.40099, 2018

The European Commission addressed Google's conduct involving Android, including restrictions connected with application distribution and mobile ecosystems.

Relevance

Autonomous economies will increasingly operate through ecosystems rather than individual products.

A single AI ecosystem could control:

Operating system → AI assistant → app marketplace → payment → identity → data → cloud → advertising

Competition problems can consequently arise from vertical integration across an ecosystem.

Institutional consequence

Competition authorities may need specialized ecosystem divisions capable of analysing:

• tying;
• interoperability;
• default settings;
• access restrictions;
• API conditions;
• data portability;
• switching costs; and
• ecosystem foreclosure.

7. Case 4: United States v Google — Search

The US litigation concerning Google's general search business illustrates the difficulty of applying conventional monopolization principles to a highly data-driven digital ecosystem. The case focused on alleged exclusionary agreements and distribution practices.

Relevance to autonomous economies

Distribution may become even more important when autonomous agents select services automatically.

Imagine millions of AI purchasing agents being programmed to use one particular:

• search engine;
• payment system;
• cloud provider;
• marketplace; or
• identity infrastructure.

Even if alternative providers technically exist, autonomous defaults may prevent effective competition.

Future institutional response

Competition authorities may therefore need to supervise:

• AI-agent default arrangements;
• choice architecture;
• interoperability;
• switching mechanisms;
• agent-access restrictions; and
• contractual restrictions affecting autonomous purchasing.

This moves competition policy from merely asking "Who has market power?" to also asking "Who controls the decision architecture through which autonomous agents enter markets?"

8. Case 5: Epic Games v Apple

Epic Games, Inc. v Apple Inc., 67 F.4th 946 (9th Cir. 2023)

The litigation concerning Apple's App Store ecosystem addressed issues involving platform rules, distribution and payment mechanisms.

Relevance

The autonomous economy may depend heavily upon:

• AI-agent marketplaces;
• agent-to-agent payment systems;
• model marketplaces;
• autonomous software stores; and
• API marketplaces.

A platform that controls access to an autonomous ecosystem may become a gatekeeper for machine commerce.

Institutional implication

Future competition authorities may need powers concerning:

Interoperability

Agents should potentially be able to interact with competing systems.

Data portability

Users and autonomous agents may need to transfer relevant data and preferences.

Payment neutrality

Platforms may not necessarily be able to force all autonomous transactions through their own payment infrastructure.

API access

Independent AI agents may require access to APIs on reasonable terms.

9. Case 6: FTC v Qualcomm

FTC v Qualcomm Inc., 969 F.3d 974 (9th Cir. 2020)

The case concerned competition issues involving technological standards, patent licensing and chipset markets.

Relevance to autonomous economies

Autonomous economies will depend heavily upon technical standards.

Examples include:

• AI-agent communication protocols;
• identity standards;
• blockchain interoperability;
• autonomous vehicle communication;
• machine-payment protocols;
• cloud interoperability;
• robotic operating systems.

Control over an essential technological layer can therefore produce competitive advantages throughout an ecosystem.

Institutional implication

Competition authorities may require technical-standard expertise capable of evaluating:

• interoperability;
• licensing;
• standard-essential technologies;
• API access;
• technical lock-in; and
• exclusionary technical specifications.

10. Case 7: Ohio v American Express

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

The US Supreme Court considered the economics of a two-sided transaction platform.

Importance for autonomous economies

Autonomous markets will frequently be multi-sided.

For example:

AI Agent ↔ Marketplace ↔ Seller

or:

Autonomous Vehicle ↔ Mobility Platform ↔ Passenger

or:

AI Buyer ↔ Payment Network ↔ Financial Institution

Competition authorities must therefore examine effects across multiple sides of the platform.

Institutional consequence

Future authorities may need sophisticated:

• platform economics;
• network-effects analysis;
• multi-sided market modelling; and
• algorithmic market simulation.

11. Case 8: Amazon Marketplace Competition Proceedings

The FTC's 2023 action against Amazon alleged that Amazon used interconnected strategies to maintain monopoly power, including conduct affecting sellers and competing services. These are allegations in ongoing litigation rather than a final judicial finding of infringement.

Relevance

Amazon demonstrates how a modern platform can simultaneously function as:

• marketplace;
• retailer;
• logistics provider;
• advertising platform;
• data intermediary; and
• infrastructure provider.

Autonomous economies may magnify this phenomenon.

An AI company could potentially become simultaneously:

AI model provider + agent marketplace + cloud provider + payment provider + data intermediary + transaction platform.

Institutional lesson

Competition authorities may need conglomerate ecosystem analysis rather than analysing each service in isolation.

12. Case 9: Google Ad-Tech Investigation

The UK CMA's continuing Google ad-tech investigation illustrates another institutional development. The CMA's 2024 statement of objections provisionally alleged that Google's conduct could restrict competition by favouring its own ad-tech services; the investigation remains distinct from a final infringement decision.

Relevance

Ad-tech is already a highly automated environment involving:

• algorithms;
• auctions;
• automated bidding;
• data;
• matching;
• ranking; and
• machine-generated decisions.

Autonomous economies could extend this model to almost every commercial sector.

Institutional implication

Competition authorities may need real-time or near-real-time market surveillance systems capable of detecting:

• coordinated pricing;
• exclusionary algorithms;
• discriminatory access;
• suspicious bidding patterns;
• self-preferencing;
• market manipulation; and
• algorithmic switching barriers.

13. Future Institutional Models

Model I — AI-Enabled Competition Authority

The traditional competition authority would remain legally responsible for decisions but use AI systems for:

• market surveillance;
• anomaly detection;
• economic modelling;
• document analysis;
• merger screening;
• algorithmic pattern recognition; and
• prediction of possible competition risks.

Structure

Human Authority

↓

AI Monitoring Division

↓

Economic Analysis Division

↓

Algorithmic Audit Division

↓

Legal Enforcement Division

The AI system would assist rather than replace legally accountable decision-makers.

14. Model II — Algorithmic Audit Authority

A specialized Algorithmic Competition Audit Office could examine important algorithms used by dominant firms.

It could test:

• pricing;
• ranking;
• recommendations;
• allocation;
• advertising;
• matching;
• procurement;
• credit;
• search;
• autonomous negotiation.

Possible powers

The authority could require:

• algorithmic documentation;
• model-risk assessments;
• audit logs;
• controlled testing;
• data-access protocols;
• explanations of material system changes.

This would transform competition enforcement from purely retrospective investigation into preventive technical supervision.

15. Model III — Digital Ecosystem Authority

Some autonomous economies may be too interconnected for traditional product-market regulation.

A specialized ecosystem authority could monitor:

1. AI ecosystems;
2. cloud ecosystems;
3. digital identity systems;
4. payment networks;
5. autonomous marketplaces;
6. data ecosystems;
7. robotics platforms.

The UK's emerging institutional model is instructive. Under its new digital-markets regime, the CMA has Strategic Market Status powers, and in October 2025 it designated Google as having SMS in general search and search advertising.

This represents movement toward ex-ante supervision of strategically important digital ecosystems.

16. Model IV — Distributed Competition Governance

Autonomous economies may operate internationally and across decentralized networks.

A single national authority may therefore be unable to monitor the entire market.

A future model could involve:

National Competition Authorities

International Competition Network

Technology Regulators

Data Protection Authorities

Financial Regulators

Cybersecurity Authorities

Technical Standards Bodies

This would create a networked competition-governance architecture.

17. Model V — Regulatory Sandbox Authority

Authorities could create controlled environments in which autonomous economic systems are tested before large-scale deployment.

For example:

Stage 1

AI-agent system enters sandbox.

Stage 2

Competition risks are assessed.

Stage 3

Interoperability is tested.

Stage 4

Pricing and coordination behaviour is simulated.

Stage 5

Market-access safeguards are established.

Stage 6

System receives conditional market authorization.

This model would be particularly relevant for:

• autonomous financial markets;
• AI procurement systems;
• autonomous vehicle networks;
• energy markets;
• machine-to-machine commerce.

18. Model VI — Continuous Competition Monitoring

Traditional enforcement generally follows:

Conduct → Investigation → Decision → Remedy

Autonomous economies may require:

Market Entry → Continuous Monitoring → Algorithmic Testing → Risk Detection → Intervention

This resembles financial-market supervision more than traditional antitrust enforcement.

The ACCC's recent digital-platform work provides an example of this monitoring-oriented direction: its 2025 report recommended that the ACCC retain a monitoring function for emerging digital technologies, including cloud computing and generative AI, and supported a permanent whole-of-government digital-regulation forum.

19. Institutional Design Principles

A. Technological neutrality

Competition authorities should regulate competitive effects, not particular technologies.

The same principles should apply whether market power is exercised through:

• AI;
• blockchain;
• cloud;
• robotics;
• quantum computing;
• traditional software.

B. Human accountability

Autonomous decision-making should not create an accountability vacuum.

A company should not be able to argue:

"The algorithm did it."

Competition law should continue to identify the undertaking responsible for deploying, controlling or materially benefiting from the system.

C. Auditability

Important autonomous systems should maintain sufficient records to allow regulators to reconstruct:

• decisions;
• model versions;
• data inputs;
• material changes;
• pricing decisions;
• transactions;
• access decisions.

D. Explainability proportional to competition risk

Not every algorithm requires complete disclosure.

However, systems capable of affecting competition significantly may require enhanced regulatory transparency.

20. Algorithmic Collusion and Autonomous Agents

One of the most difficult future problems is machine-to-machine coordination.

Suppose four autonomous pricing agents repeatedly interact.

They may independently discover that maintaining high prices maximizes long-term profits.

No human communicates.

The question becomes:

Can competition law intervene when coordination emerges without an express agreement?

Future institutional models may require authorities to distinguish:

Legitimate parallel behaviour

from

Algorithmically facilitated coordination

and from

Deliberately designed coordination mechanisms.

This requires economists, computer scientists and lawyers to work together.

21. Data as a Source of Institutional Competition Power

Autonomous economies will make data even more important.

A dominant firm may possess:

• transaction data;
• behavioural data;
• training data;
• real-time market data;
• agent interaction data;
• logistics data.

Data advantages can create:

Data → Better AI → More Users → More Data → Better AI

This produces a feedback loop.

Competition authorities therefore may need to investigate data-based barriers to entry, not merely traditional market shares.

22. Autonomous Mergers

Traditional merger control examines whether two firms combining will substantially lessen competition.

Autonomous economies introduce additional questions.

For example:

AI Model A + Cloud Platform B

could produce:

• greater computing access;
• superior training capacity;
• integrated distribution;
• exclusive data;
• autonomous-agent access.

Authorities may therefore need to assess ecosystem accumulation, not merely the immediate overlap between the merging firms.

23. Structural Remedies in Autonomous Economies

Traditional remedies include:

• fines;
• behavioural commitments;
• licensing;
• divestiture.

Future autonomous economies may require:

Interoperability remedies

Competitors receive technical access.

Data-portability remedies

Users can move relevant information.

API-access remedies

Competitors receive reasonable technical access.

Algorithmic neutrality remedies

Dominant platforms cannot discriminate against competitors through ranking systems.

Functional separation

A platform may be required to separate:

• infrastructure;
• marketplace;
• data;
• ranking;
• advertising.

Independent technical monitoring

A neutral monitor could continuously examine compliance.

24. Institutional Independence

Future competition authorities should be protected against:

• political interference;
• industry capture;
• technological dependency;
• excessive dependence on regulated firms for technical expertise.

Institutional independence becomes especially important because autonomous economies may be dominated by firms possessing greater technical resources than the regulator.

25. Multi-Disciplinary Competition Authorities

A future authority may need five major professional groups:

A traditional competition authority staffed almost entirely by lawyers and economists may therefore be insufficient for highly autonomous markets.

26. International Cooperation

Autonomous economic networks can operate globally.

An AI agent in India could:

1. obtain data from Europe;
2. purchase cloud services in the US;
3. transact through Singapore;
4. contract with an autonomous supplier in Japan; and
5. distribute goods through a global platform.

Consequently, competition authorities will increasingly require:

• evidence-sharing mechanisms;
• coordinated investigations;
• common technical standards;
• compatible merger-review procedures;
• cross-border algorithmic audits.

27. Due Process and Regulatory Limits

Institutional innovation must not eliminate procedural safeguards.

Autonomous-economy regulation should preserve:

• notice;
• hearing rights;
• confidentiality;
• judicial review;
• evidentiary standards;
• proportionality;
• appeal rights;
• protection of trade secrets.

A competition authority should not become an uncontrolled technological regulator merely because markets are becoming autonomous.

28. Proposed Future Institutional Architecture

A comprehensive institutional model could look like this:

                 NATIONAL COMPETITION AUTHORITY                            │             ┌──────────────┼──────────────┐             │              │              │       Legal Division   Economics      AI/Tech Division                            │              │                            │        Algorithmic Audit                            │              │                     Market Intelligence  │                            │              │             ┌──────────────┴──────────────┘             │       Ecosystem Supervision             │      ┌──────┼──────┬────────┐      │      │      │        │     AI    Cloud   Data   Autonomous  Markets  Markets Markets  Agents      │      │      │        │      └──────┼──────┴────────┘             │       Continuous Monitoring             │       Risk-Based Intervention             │       Human Decision-Maker             │       Judicial Review

29. Core Legal Questions for Future Autonomous Economies

Competition authorities will increasingly have to answer:

1. Who is the undertaking?

The AI itself ordinarily cannot simply replace the legal responsibility of the human or corporate actor behind it.

2. Who controls the algorithm?

Ownership, deployment, modification and economic benefit may need separate examination.

3. Can autonomous coordination constitute prohibited coordination?

This will become increasingly important where algorithms interact repeatedly.

4. Who owns economically important data?

Data access may determine competitive opportunity.

5. Can autonomous agents switch platforms?

Switching capability may become a major measure of contestability.

6. Can competing AI agents interoperate?

Interoperability may become analogous to access to essential infrastructure in particular circumstances.

7. Can a dominant AI platform self-preference?

The Google Shopping experience demonstrates why ranking and self-preferencing can become central competition questions.

8. How should remedies operate?

Static remedies may be ineffective if algorithms evolve continuously.

30. Six Major Institutional Lessons from the Case Law

31. Future Model: From Competition Authority to Competition Infrastructure

The most significant institutional transformation may be conceptual.

The competition authority of the future may not simply be:

an institution that prosecutes antitrust violations.

It may become:

a permanent competition infrastructure that continuously observes markets, audits strategically important algorithms, monitors ecosystem access, evaluates structural changes and intervenes when competitive conditions materially deteriorate.

This does not mean replacing traditional enforcement. Rather, it adds an ex-ante and technologically sophisticated layer to existing competition law.

32. Conclusion

Autonomous economies challenge the traditional assumption that markets consist primarily of identifiable human-controlled firms making relatively observable decisions.

The next generation of competition law may therefore require a transition:

Firm-centred → Ecosystem-centred

Human decision-centred → Human-and-machine decision-centred

Ex-post enforcement → Ex-post + ex-ante supervision

Periodic investigation → Continuous monitoring

Legal/economic expertise → Legal + economic + computational expertise

National enforcement → Networked international enforcement

The Google Shopping, Android, Epic Games, Qualcomm, American Express, Google Search, Amazon and emerging ad-tech experiences demonstrate the increasing importance of platform architecture, data, interoperability, self-preferencing, technological infrastructure and ecosystem power. Current regulatory developments reinforce this trajectory: the UK has moved toward Strategic Market Status supervision, while the ACCC has explicitly examined cloud and generative-AI competition issues and recommended continuing monitoring of emerging technologies.

Ultimately, the institutional model for autonomous economies should preserve the foundational objectives of competition law—contestability, consumer choice, innovation, market access and prevention of exclusionary power—while developing the technical capacity necessary to understand markets in which economically consequential decisions increasingly occur through autonomous computational systems.