Competition Law and Institutional Transformation of Antitrust Authorities in AGI Economies

1. Introduction

The emergence of Artificial General Intelligence (AGI) economies could fundamentally change how competition authorities detect, investigate, and remedy anticompetitive conduct. An AGI economy may involve highly autonomous AI systems capable of performing complex research, pricing, contracting, product development, investment allocation, logistics, procurement, and strategic decision-making across multiple markets.

This creates an institutional question for competition law:

Can traditional antitrust authorities, designed primarily to investigate human firms and relatively observable markets, effectively regulate markets in which autonomous AI systems continuously make commercial decisions, control infrastructure, generate innovations, and interact with one another?

The issue is not simply whether competition statutes should change. It concerns the institutional transformation of competition authorities themselves—their expertise, investigative methods, technical infrastructure, organizational structure, evidence-gathering powers, merger-review capabilities, market-monitoring systems, and remedies.

AGI does not automatically create an antitrust violation. The central competition-law concern arises when autonomous systems operate within markets characterized by concentration, network effects, data advantages, infrastructure dependencies, interoperability restrictions, algorithmic coordination, or exclusionary conduct.

2. Meaning of Institutional Transformation

Institutional transformation means the adaptation of competition authorities from conventional enforcement institutions into organizations capable of supervising increasingly autonomous and technologically complex markets.

A traditional authority generally relies upon:

complaints;

documents and emails;

witness testimony;

economic evidence;

market-definition analysis;

company interviews;

transaction documents;

pricing information; and

retrospective investigation.

An AGI economy may require additional capabilities:

continuous algorithmic monitoring;

AI-assisted market analysis;

model auditing;

technical inspection of AI systems;

monitoring of autonomous agents;

computational competition analysis;

real-time detection of coordinated behaviour;

analysis of machine-generated contracts;

monitoring of APIs and interoperability;

evaluation of AI training-data dependencies;

monitoring of compute and cloud concentration;

predictive merger screening; and

specialized technological remedies.

Thus, institutional transformation concerns how antitrust authorities themselves must evolve as market structures evolve.

3. Why AGI Creates a New Institutional Challenge

AGI could alter competition along several dimensions.

A. Autonomous commercial decision-making

AI agents could independently:

negotiate contracts;

determine prices;

purchase inputs;

select suppliers;

allocate advertising;

manage inventories;

develop products;

conduct research; and

enter strategic partnerships.

The traditional distinction between a firm's human decision-maker and its technological instrument may therefore become less clear.

B. Speed

Human enforcement may take months or years.

An autonomous system could alter market behaviour in seconds.

Consequently, retrospective enforcement may become less effective where a temporary exclusionary strategy can cause irreversible competitive damage.

C. Complexity

AGI systems may contain enormous numbers of interacting variables.

A competition authority may therefore need to understand not merely what a company did but:

what model made the decision;

what data influenced the model;

what objective function was optimized;

what constraints were imposed;

what external systems were connected;

whether the system learned from competitors' behaviour; and

whether the resulting conduct was predictable.

D. Concentration of enabling infrastructure

AGI competition may depend upon scarce inputs such as:

advanced semiconductors;

computing capacity;

cloud infrastructure;

specialized data;

foundation models;

AI talent;

energy;

distribution platforms; and

proprietary technical interfaces.

Competition authorities therefore may increasingly need to examine vertical layers of the AI economy simultaneously.

4. Institutional Transformation of Competition Authorities

A. From retrospective enforcement to continuous market monitoring

Traditional antitrust enforcement is substantially retrospective.

In an AGI economy, authorities may increasingly require continuous market intelligence.

For example, an authority could monitor:

changes in market concentration;

algorithmic pricing patterns;

access restrictions;

API availability;

switching costs;

compute allocation;

acquisition patterns;

licensing conditions;

interoperability;

algorithmic coordination indicators; and

emerging competitors.

This does not mean that every market should be subjected to permanent government supervision. Rather, high-risk digital markets may require earlier detection mechanisms.

5. AI-Assisted Antitrust Investigation

Competition authorities themselves may use AI to process enormous quantities of evidence.

AI could help identify:

suspicious communications;

unusual pricing patterns;

coordinated bidding;

exclusionary contractual provisions;

discriminatory access;

acquisition patterns;

common ownership relationships;

parallel algorithmic behaviour; and

strategically significant technological dependencies.

This could transform the authority from an institution primarily dependent upon manually collected evidence into a computational enforcement institution.

However, AI-assisted enforcement creates its own risks.

An authority must be able to explain:

what data the system used;

how the system reached its analytical conclusion;

whether false positives occurred;

whether the model was biased;

whether defendants can challenge the methodology; and

whether confidential business information was properly protected.

6. Specialized AGI Competition Units

Competition authorities may require specialized units covering:

1. AI economics

Economists capable of analysing:

algorithmic pricing;

dynamic competition;

network effects;

innovation markets;

multi-sided platforms; and

AI-driven market structure.

2. AI engineering

Technical experts capable of understanding:

model architecture;

training systems;

APIs;

inference infrastructure;

agentic systems;

model deployment; and

interoperability.

3. Data economics

Experts could assess:

data accumulation;

data portability;

data advantages;

data exclusivity;

data interoperability; and

data-network effects.

4. Computational forensics

This would involve examining:

model logs;

source code where legally obtainable;

system prompts;

model outputs;

decision histories;

APIs;

training datasets; and

autonomous-agent interactions.

7. Institutional Transformation and Merger Control

AGI makes merger control particularly important.

A traditional merger analysis may focus upon existing market shares.

But an AGI transaction may involve an apparently small company possessing:

an important research team;

unique training data;

a specialized model;

a critical API;

an important algorithm;

an emerging technology; or

a potentially disruptive innovation.

Consequently, competition authorities may need to examine innovation pipelines and potential competition, not merely current turnover.

8. Case Law

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

The Microsoft litigation remains important for AGI institutional transformation because it demonstrates the importance of understanding technological ecosystems rather than isolated products.

Microsoft was found to have engaged in exclusionary conduct concerning Internet browsers and competing technologies.

The case demonstrated that competition authorities must understand:

technological interoperability;

platform control;

distribution;

developer relationships;

technical standards; and

leverage between adjacent markets.

Relevance to AGI

An AGI authority may similarly need to investigate relationships among:

foundation models → cloud infrastructure → APIs → applications → distribution platforms.

The institutional lesson is that antitrust agencies require deep technical expertise when investigating platform-based ecosystems.

9. Case 2: Microsoft Corp. v Commission — T-201/04

The European Commission's Microsoft decision concerned interoperability information and tying.

The case demonstrated that competition law can become closely connected with technical information and interoperability.

AGI relevance

AGI markets could involve:

model interoperability;

agent interoperability;

API access;

data portability;

compatibility standards; and

access to technical interfaces.

An authority therefore may require personnel capable of determining whether technical restrictions merely protect legitimate innovation or instead exclude competitors.

The case supports an important institutional proposition:

Competition authorities cannot effectively regulate technologically complex markets without technological competence.

10. Case 3: IMS Health v NDC Health — C-418/01

The Court of Justice examined refusal to provide access to a copyrighted data structure.

The judgment established demanding conditions under which refusal to license intellectual property can constitute an abuse of dominance.

AGI relevance

AGI ecosystems may contain strategically important:

datasets;

data architectures;

APIs;

model interfaces;

technical standards; and

interoperability structures.

Authorities must therefore develop expertise capable of distinguishing:

legitimate protection of intellectual property; from

exclusionary control over an indispensable competitive input.

This requires institutional expertise in both competition law and technology law.

11. Case 4: Oscar Bronner v Mediaprint — C-7/97

Bronner concerned access to a newspaper distribution system.

The Court adopted a strict approach toward compulsory access to infrastructure under the essential-facilities doctrine.

AGI relevance

AGI economies may contain infrastructure that resembles essential facilities:

cloud computing;

specialized AI chips;

model-hosting infrastructure;

data-access systems;

AI marketplaces;

critical APIs.

But competition authorities must not automatically treat every technologically important resource as an essential facility.

Institutional transformation therefore requires authorities to develop sophisticated economic tests concerning:

indispensability;

alternatives;

duplication;

investment incentives;

foreclosure; and

downstream competition.

12. Case 5: Commercial Solvents v Commission — Joined Cases 6/73 and 7/73

Commercial Solvents established important principles concerning refusal to supply and downstream foreclosure.

A dominant undertaking controlling an upstream input could not simply use that position to eliminate competition in a downstream market.

AGI relevance

Imagine an enterprise controlling:

critical AI compute;

a foundation model;

an essential data resource;

while also competing downstream through AI applications.

The authority would need to investigate whether the firm is using upstream control to disadvantage downstream competitors.

This requires vertical-market institutional analysis, rather than simply looking at one product market.

13. Case 6: Google Shopping — Commission Decision AT.39740

The Google Shopping case involved the treatment of comparison-shopping services in Google's search results.

The case illustrates the importance of examining ranking systems and platform design as potential instruments of competitive advantage.

AGI relevance

In an AGI economy, autonomous systems may determine:

which application is recommended;

which supplier is selected;

which agent receives visibility;

which model is called;

which service is prioritized; and

which information is presented to users.

Competition authorities may therefore need to understand algorithmic ranking and recommendation architecture.

Institutionally, this means antitrust authorities may require algorithmic auditing capabilities.

14. Case 7: Google Android — Commission Decision AT.40099

The Google Android case concerned restrictions involving mobile operating systems, applications, search and distribution.

The case is particularly relevant because it illustrates how control of one technological layer can influence competition in adjacent layers.

AGI relevance

An AGI ecosystem could involve:

operating system → device → cloud → foundation model → assistant → application marketplace.

A dominant firm controlling several layers may have opportunities to leverage its position.

Authorities therefore may need organizational structures capable of analysing vertical technological ecosystems rather than traditional standalone markets.

15. Case 8: Facebook/Meta Data Case — Bundeskartellamt

The German Facebook proceedings addressed the relationship between market power and extensive collection/combination of user data.

The case demonstrated that data practices can become relevant to competition analysis where a dominant platform's position enables it to impose or implement conditions affecting users and competitors.

AGI relevance

Data may become even more important in AGI markets because competitive advantages can arise from:

training data;

interaction data;

behavioural information;

enterprise data;

feedback loops;

proprietary datasets.

Competition authorities may consequently require data economists, data scientists and privacy-competition specialists.

16. Institutional Transformation and Algorithmic Collusion

One of the most difficult AGI-related issues is potential algorithmic coordination.

Traditional cartel enforcement often looks for:

communications;

meetings;

agreements;

instructions;

price exchanges.

Autonomous systems could potentially produce parallel outcomes without conventional human communication.

For example, competing AI agents could learn that aggressive price competition reduces profits and independently adopt strategies that produce stable high prices.

This creates an institutional problem:

How should an authority distinguish legitimate autonomous adaptation from unlawful coordination?

Competition authorities may therefore require:

algorithmic testing;

simulation;

code and model examination;

behavioural experiments;

economic modelling;

audit trails; and

real-time monitoring.

Importantly, parallel AI behaviour alone should not automatically be treated as proof of a cartel.

17. AGI and Evidence Collection

Traditional evidence may increasingly become insufficient.

Authorities may need access, subject to applicable procedural safeguards, to:

model-development records;

model versions;

decision logs;

API calls;

agent communications;

system instructions;

training documentation;

deployment records;

contractual restrictions;

cloud allocation records; and

algorithmic outputs.

The procedural challenge is significant because AI systems may generate millions of decisions.

Authorities therefore need technology capable of identifying legally significant evidence within enormous computational datasets.

18. AGI and Merger Screening

Institutional transformation may be particularly significant in merger control.

Authorities may need to identify:

Killer acquisitions

An incumbent could acquire a small AI firm before the latter becomes a meaningful competitor.

Capability acquisitions

A transaction may involve acquiring:

AI researchers;

algorithms;

datasets;

compute technology;

specialized models.

Ecosystem acquisitions

An acquisition could strengthen control over an entire technological stack.

Vertical acquisitions

A cloud provider could acquire a model developer, or a model developer could acquire a downstream application platform.

The relevant question may therefore be:

What competitive capability is being acquired, rather than merely what current revenue is being acquired?

19. From Market Definition to Capability Mapping

Traditional market definition remains important, but AGI may require complementary analysis.

Authorities may map:

compute;

models;

data;

talent;

APIs;

applications;

distribution;

cloud;

infrastructure;

users;

developers.

This produces a capability map.

A firm with modest market share in an application market could nevertheless possess significant strategic power if it controls an essential technological layer.

20. Institutional Transformation and Remedies

AGI competition cases may require remedies different from conventional structural remedies.

Possible remedies include:

A. Interoperability

Require systems to communicate with competing services.

B. Data portability

Allow users or businesses to transfer relevant data.

C. API access

Prevent unjustified discriminatory access to essential interfaces.

D. Non-discrimination obligations

Require dominant platforms to apply access conditions consistently.

E. Separation remedies

In exceptional circumstances, separate infrastructure from downstream competitive activities.

F. Transparency obligations

Require explanation of certain ranking or access mechanisms.

G. Monitoring trustees

Specialized technical monitoring could verify compliance.

H. Merger conditions

Transactions could be approved subject to obligations protecting:

interoperability;

access;

innovation;

data mobility; and

non-discrimination.

21. The Role of Competition Authorities in Innovation Policy

AGI creates tension between competition enforcement and innovation incentives.

Overly aggressive intervention may potentially discourage:

AI investment;

research;

infrastructure development;

intellectual-property investment; and

risky technological experimentation.

Insufficient intervention, however, could allow incumbents to entrench themselves before competitors can emerge.

Authorities therefore need institutional capacity to evaluate dynamic competition.

The question is not simply:

“Is the firm large?”

It is also:

“Does the firm's conduct reduce the ability of future technologies and competitors to emerge?”

22. Competition Authority as a Technological Institution

In an AGI economy, competition authorities may increasingly resemble multidisciplinary institutions containing:

This would represent a significant institutional transformation.

23. Institutional Independence and AGI

Technological expertise alone is insufficient.

Competition authorities must also maintain:

legal independence;

procedural fairness;

transparent decision-making;

confidentiality;

judicial review;

protection against political interference;

accountability; and

consistency.

An authority that relies heavily on AI systems for enforcement must also ensure that AI does not replace legally accountable human decision-making.

AI may assist the authority, but legal responsibility should remain institutionally identifiable.

24. Risk of Regulatory Capture

AGI may produce a particularly difficult form of regulatory dependence.

Large technology firms may possess:

more computing capacity;

more technical experts;

more proprietary data;

more sophisticated AI systems;

greater cybersecurity capabilities.

If authorities lack equivalent technical capacity, an information asymmetry may emerge between regulator and regulated firm.

Therefore institutional transformation should include investment in the authority's own technological infrastructure and independent expertise.

25. International Cooperation

AGI markets will frequently be transnational.

A foundation model could be:

developed in one country;

trained using globally sourced data;

hosted on infrastructure in another;

distributed through international platforms; and

deployed worldwide.

Competition authorities will therefore need greater cooperation concerning:

merger investigations;

evidence;

algorithmic conduct;

digital markets;

cross-border acquisitions;

remedies;

data;

cloud infrastructure; and

international cartels.

International cooperation becomes especially important where one authority cannot observe the entire technological ecosystem.

26. Institutional Transformation and India

For India, the Competition Commission of India may face several AGI-related institutional questions.

The Competition Act, 2002 already provides the legal framework for addressing:

anti-competitive agreements under Section 3;

abuse of dominant position under Section 4;

combinations under Sections 5 and 6;

factors relevant to combination assessment under Section 20(4).

Several existing factors become particularly important for AGI markets, including:

barriers to entry;

economic power;

technological advantages;

consumer benefits;

innovation;

market access;

vertical integration;

control of infrastructure; and

removal of effective competition.

The institutional challenge would therefore be less about abandoning traditional competition law and more about developing the technical capacity necessary to apply existing principles to technologically transformed markets.

27. Core Institutional Risks

AGI economies could produce several risks for competition authorities:

1. Information asymmetry

Companies may understand their systems better than regulators.

2. Enforcement latency

Markets can change faster than investigations.

3. Technical complexity

Traditional investigative techniques may be insufficient.

4. Evidence overload

AI systems can produce enormous datasets.

5. Innovation uncertainty

Future competitive effects may be difficult to predict.

6. Cross-market leverage

Power can move rapidly from one technological layer to another.

7. Regulatory fragmentation

Competition, data, AI, financial and sectoral regulators may investigate overlapping conduct.

8. Regulatory capture

Highly concentrated technical expertise may create dependence on dominant firms.

28. A New Institutional Model

A mature AGI competition authority could potentially operate through several complementary functions:

Market Observatory → AI Risk Detection → Economic Analysis → Technical Investigation → Legal Assessment → Merger/Conduct Review → Remedy Design → Continuous Monitoring

This would transform antitrust from a predominantly reactive system into a combination of:

enforcement;

market monitoring;

technological expertise;

economic analysis; and

institutional learning.

29. Key Principles Emerging from the Case Law

The eight cases collectively illustrate several principles relevant to AGI institutional transformation:

Microsoft — competition authorities must understand technological ecosystems.

Microsoft v Commission — interoperability can become a competition issue.

IMS Health — access to protected technological resources requires careful legal analysis.

Bronner — essential-facility intervention requires demanding conditions.

Commercial Solvents — upstream control can create downstream foreclosure concerns.

Google Shopping — algorithmic ranking can affect competitive conditions.

Google Android — control across technological layers can create leverage.

Facebook/Meta — data concentration can intersect with dominance analysis.

Together, these cases demonstrate that competition enforcement is increasingly dependent upon technical, economic and institutional expertise.

30. Conclusion

Institutional transformation of antitrust authorities is likely to become a central competition-policy issue in AGI economies.

The fundamental challenge is not that AGI necessarily makes existing competition law obsolete. Rather, AGI may make traditional institutional capabilities insufficient for applying that law effectively.

Competition authorities may increasingly need to transform themselves from organizations primarily focused on conventional corporate conduct into multidisciplinary technological-economic institutions capable of understanding:

autonomous agents;

algorithmic decision-making;

AI infrastructure;

cloud and compute concentration;

data ecosystems;

interoperability;

algorithmic coordination;

innovation pipelines;

technological acquisitions;

platform ecosystems; and

rapidly changing competitive conditions.

The central legal principle should remain that technology or AGI capability itself is not unlawful. Competition law intervenes where market power is used in ways that produce legally relevant exclusion, coordination, exploitation, foreclosure, or anticompetitive effects.

Accordingly, the institutional transformation of antitrust authorities should combine technical competence, economic sophistication, procedural safeguards, international cooperation and legal accountability. The future effectiveness of competition law in AGI economies may depend as much on the institutional capacity of enforcement agencies as on the substantive rules contained in competition statutes.