AI Border Control Systems and Digital Market Partitioning

1. Introduction

AI border control systems include automated passport gates, biometric identification, facial-recognition systems, risk-scoring engines, passenger pre-screening platforms, automated visa systems, customs-risk analytics, and interoperable databases used by border authorities.

These systems can create competition concerns when a dominant technology provider, infrastructure operator, or vertically integrated ecosystem uses control over AI models, biometric databases, APIs, identity infrastructure, cloud services, software interfaces, or certification systems to divide markets geographically or technically.

Digital market partitioning occurs when competition is restricted by artificially separating customers, suppliers, data, technologies, or transactions according to territory, platform, user category, technical environment, or access conditions.

The legal analysis generally involves:

• abuse of dominance;
• refusal of access;
• discriminatory interoperability;
• exclusionary technical standards;
• geographic restrictions;
• tying and bundling;
• self-preferencing;
• discriminatory access to data;
• interoperability restrictions;
• exclusive procurement;
• foreclosure of competing AI or identity-service providers; and
• agreements that partition markets.

There is not yet a large body of reported competition litigation specifically concerning AI-powered border-control systems. Therefore, established competition cases concerning essential facilities, interoperability, territorial restrictions, digital platforms and market foreclosure provide the principal legal analogies.

2. Meaning of Digital Market Partitioning

Digital market partitioning can operate through several mechanisms.

A. Geographic partitioning

An AI border system may be configured so that:

• biometric services work only in particular jurisdictions;
• competing identity providers cannot operate across borders;
• software licenses are geographically restricted;
• passenger or customs data cannot be transferred to competing systems;
• APIs are available only to selected national operators.

B. Technical partitioning

A dominant system may make its technology interoperable with its own products but difficult to connect with competing systems.

Examples include:

• undocumented APIs;
• incompatible data formats;
• proprietary biometric templates;
• closed identity protocols;
• restrictive software-development kits;
• delayed certification of competing systems.

C. Data partitioning

A dominant border technology operator may control:

• facial-recognition datasets;
• passenger-information databases;
• watch-list interfaces;
• biometric identity databases;
• risk-scoring information;
• historical customs information.

If competitors cannot obtain reasonably necessary access, the data advantage may reinforce market power.

D. Platform partitioning

A border-control platform may give preferential treatment to:

• its own identity-verification services;
• its own cloud infrastructure;
• affiliated cybersecurity services;
• affiliated analytics products.

This can transform an initially legitimate border-security infrastructure into a mechanism for excluding competitors.

3. Competition-Law Framework

3.1 Abuse of Dominance

Where an undertaking possesses substantial market power in an essential digital infrastructure market, competition law may prohibit conduct that excludes competitors without sufficient objective justification.

Potentially relevant conduct includes:

1. refusal to supply;
2. discriminatory access;
3. tying;
4. bundling;
5. exclusive dealing;
6. technical degradation;
7. interoperability restrictions;
8. discriminatory pricing;
9. territorial restrictions; and
10. leveraging dominance into adjacent AI markets.

4. AI Border Control as a Potential Essential Facility

The first question is whether the relevant border-control infrastructure is genuinely indispensable.

Possible examples could include:

• a single national biometric gateway;
• a government-certified identity interface;
• an exclusive passenger-information database;
• a unique border-risk database;
• a mandatory interoperability gateway.

However, mere usefulness is not enough.

Competition law generally requires a considerably stronger showing of indispensability before imposing a compulsory-access obligation.

5. Refusal of Access and Interoperability

Suppose a private company operates the AI platform used at an airport.

It might control the interface through which airlines, security providers and alternative identity-verification providers access the border system.

If the operator:

• refuses API access;
• supplies incomplete technical documentation;
• provides inferior access to rivals;
• imposes unreasonable technical conditions; or
• requires competitors to purchase unrelated services,

competition concerns can arise.

The critical distinction is between:

legitimate security requirements

and

competition-restricting technical exclusion disguised as security requirements.

Border security may constitute an important objective justification, but the restriction should ordinarily be connected to that objective and proportionate to it.

6. Territorial Restrictions and Digital Market Partitioning

Traditional competition law has long been concerned with restrictions that divide markets according to territory.

Digital systems can achieve the same result without an explicit territorial clause.

For example:

Country A → AI provider A → closed API → Country A customers

while

Country B → AI provider B → separate database → Country B customers

may create technological barriers preventing cross-border competition.

Such partitioning can reduce:

• consumer choice;
• cross-border service provision;
• economies of scale;
• interoperability;
• innovation; and
• entry by competing technology suppliers.

7. Important Case Laws

Case 1 — United Brands v Commission

United Brands Company and United Brands Continentaal BV v Commission, Case 27/76

The Court of Justice examined restrictions capable of maintaining the separation of national markets.

Principle

The case is important for understanding the competition-law concern with artificial territorial segmentation.

A dominant undertaking cannot necessarily use contractual or commercial mechanisms to prevent effective competition between national markets.

Relevance to AI border systems

An AI border-control supplier could potentially partition the market through:

• country-specific licensing;
• restrictions preventing cross-border deployment;
• territorial pricing;
• separate technical environments; or
• contractual restrictions on cross-border service provision.

The existence of national security or regulatory requirements, however, may provide legitimate reasons for some differentiation.

Case 2 — Commercial Solvents v Commission

Istituto Chemioterapico Italiano S.p.A. and Commercial Solvents Corporation v Commission, Joined Cases 6/73 and 7/73

This is a foundational refusal-to-supply case.

Principle

A dominant undertaking controlling an upstream input may abuse its position when it cuts off supplies to a downstream competitor in circumstances where the input is important for downstream competition.

AI application

Consider:

AI infrastructure provider → biometric verification interface → competing border-security companies

If the infrastructure provider supplies the interface to its own downstream services but refuses access to competitors, the conduct could raise analogous foreclosure concerns.

The crucial question would be whether the refusal eliminates or substantially restricts effective competition in the downstream market.

8. Case 3 — Bronner v Mediaprint

Oscar Bronner GmbH & Co. KG v Mediaprint Zeitungs und Zeitschriftenverlag GmbH & Co. KG, Case C-7/97

Principle

The Court established a demanding framework for treating infrastructure as indispensable.

The refusal must concern a facility that is effectively indispensable, and there must generally be no realistic alternative.

Application to AI border control

This is particularly relevant.

Suppose an airport has one AI identity-verification infrastructure and an alternative provider claims:

"I cannot compete unless I receive access to the airport's biometric gateway."

Bronner indicates that competition law does not automatically require the infrastructure owner to provide access.

Questions would include:

1. Is the infrastructure genuinely indispensable?
2. Can another technically and economically viable system be created?
3. Is duplication realistically possible?
4. Would refusal eliminate effective competition?
5. Is there an objective justification?

Thus, essentiality must be demonstrated rather than presumed.

9. Case 4 — IMS Health v NDC Health

IMS Health GmbH & Co. KG v NDC Health GmbH & Co. KG, Case C-418/01

Principle

The Court considered circumstances in which refusal to license an intellectual-property-related resource could constitute an abuse.

The case established stringent conditions for compulsory access, including circumstances involving indispensability and elimination of competition.

AI application

An AI border system might contain:

• proprietary biometric architecture;
• unique data structures;
• specialized interoperability protocols;
• proprietary authentication technology.

If competitors require access to such technology to compete effectively, the IMS Health reasoning becomes relevant.

However, proprietary technology does not automatically become an essential facility.

10. Case 5 — Slovak Telekom v Commission

Slovak Telekom a.s. v European Commission, Joined Cases C-165/19 P and C-167/19 P

Principle

The case concerned exclusionary conduct involving access to infrastructure and conditions imposed on competitors.

It is important for understanding how a dominant infrastructure operator can use access conditions to restrict downstream competition.

AI border-control relevance

An AI border platform could potentially impose:

• excessive access charges;
• discriminatory technical requirements;
• inferior access;
• unnecessarily complex certification procedures;
• restrictive interoperability conditions.

For example:

Competitor A receives real-time API access, while competitor B receives delayed or technically incomplete access.

Even if formal access exists, access that is practically unusable may raise competition concerns depending on the circumstances.

11. Case 6 — Google Shopping

Google and Alphabet — Google Shopping, Case T-612/17

The EU Google Shopping litigation concerned Google's treatment of competing comparison-shopping services within its search ecosystem.

Principle

A dominant digital platform can potentially use control over an important digital gateway to advantage its own downstream service while disadvantaging competing services.

AI border-control application

Imagine a dominant border platform providing:

Identity verification + risk scoring + customs analytics + third-party AI services.

If the platform gives its own AI services preferential visibility, data access, processing speed or API functionality, while competing services receive inferior treatment, the conduct could raise self-preferencing and leveraging concerns.

The analogy is especially relevant where the border-control platform becomes a gateway through which competitors must reach customers.

12. Case 7 — Google Android

Google LLC and Alphabet Inc. v European Commission, Case T-604/18

The Android proceedings examined contractual and ecosystem restrictions involving Google's mobile ecosystem.

Principle

Competition analysis can consider how contractual restrictions operating across interconnected digital markets can reinforce an undertaking's position in adjacent markets.

AI border-control relevance

A border-control provider might control an ecosystem consisting of:

• biometric identification;
• cloud hosting;
• AI analytics;
• identity credentials;
• cybersecurity;
• airport software.

It could then condition access to the principal border platform upon purchasing additional services.

This creates a potential ecosystem foreclosure problem.

13. Case 8 — Pierre Fabre

Pierre Fabre Dermo-Cosmétique SAS, Case C-439/09

Principle

The case concerned a contractual restriction that effectively prevented distributors from selling products through the internet.

The case illustrates competition law's concern with restrictions that can substantially interfere with a distribution channel.

Digital border-control relevance

The analogy is useful where a technology provider prevents customers or authorized operators from using alternative digital channels.

Examples might include:

• prohibition on alternative identity-verification interfaces;
• restrictions on independent digital authentication;
• contractual prohibition on connecting competing AI services;
• restrictions preventing customers from using alternative cloud or analytics providers.

The legality would depend on the specific market and justification.

14. Case 9 — Magill

Radio Telefis Eireann (RTE) and Independent Television Publications Ltd (ITP) v Commission, Joined Cases C-241/91 P and C-242/91 P

Principle

Magill is a classic authority concerning compulsory access to information controlled by dominant undertakings.

The case is particularly relevant to the relationship between:

• information;
• intellectual property;
• market access; and
• downstream competition.

AI border-control relevance

AI systems depend heavily on data.

A dominant provider could control:

• historical passenger data;
• identity-verification information;
• technical datasets;
• machine-readable border information.

If that information is genuinely indispensable for competing downstream services, Magill provides an important analytical reference.

15. Case 10 — Microsoft

Microsoft Corp. v Commission, Case T-201/04

Principle

The Microsoft litigation is particularly significant for interoperability.

The case concerned Microsoft's refusal to provide interoperability information necessary for competing work-group server products.

Relevance to AI border-control systems

This is one of the most useful analogies for AI border infrastructure.

Suppose a dominant border-control system refuses to provide competitors with necessary interoperability information concerning:

• biometric interfaces;
• authentication protocols;
• passenger-information formats;
• security APIs;
• identity-token protocols.

The competition question becomes:

Is the refusal genuinely necessary for security, or does it unnecessarily preserve the incumbent's technological advantage?

A government or operator could legitimately restrict sensitive information. But the analysis may distinguish security-sensitive information from information that is unnecessarily withheld to protect commercial dominance.

16. Case 11 — MEO

MEO – Serviços de Comunicações e Multimédia SA v Autoridade da Concorrência, Case C-525/16

Principle

Discriminatory treatment by a dominant undertaking requires examination of whether the conduct places trading partners at a competitive disadvantage.

AI application

A border-control platform might charge:

• Airline A: €1 per verification
• Airline B: €1 per verification
• Rival AI provider: €10 per verification

Alternatively, it could provide superior API performance to affiliated entities.

The key competition question would involve the actual competitive effects and the circumstances of the discrimination.

17. Competition Issues Created by AI Border-Control Systems

17.1 Data monopoly

The operator with the largest border dataset may obtain significant advantages in:

• machine learning;
• identity matching;
• anomaly detection;
• fraud detection;
• risk classification.

Competitors may face a data-entry barrier that becomes progressively harder to overcome.

This can create a data-driven feedback loop:

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

17.2 Network effects

Border infrastructure can generate strong network effects.

The more airports, airlines, governments and identity systems connected to one platform, the more attractive the platform becomes.

This can produce:

Scale → interoperability → data accumulation → AI improvement → greater scale.

Competition law may therefore need to examine not merely current market share but also dynamic foreclosure.

18. Algorithmic Market Partitioning

AI can partition markets without conventional contractual restrictions.

For example, an algorithm could automatically determine:

• which suppliers can access particular users;
• which identity providers are displayed;
• which verification services are activated;
• which countries receive particular functionality;
• which customers receive preferential pricing.

This creates a new form of partitioning:

algorithmically enforced market separation.

The absence of a human instruction does not necessarily eliminate competition concerns.

The legal analysis would focus on the undertaking's conduct, control, knowledge, implementation and effects.

19. Geo-Fencing and Border Technology

Geo-fencing may be legitimate for:

• national-security requirements;
• statutory data-localization rules;
• export controls;
• privacy requirements;
• immigration restrictions;
• cybersecurity controls.

However, competition issues may arise where geo-fencing goes beyond those legitimate objectives.

For example:

A system technically capable of interoperating with competing providers is deliberately configured to prevent their operation in certain territories solely to protect an affiliated provider.

That could constitute a materially different competition problem from geo-fencing required by law.

20. Vertical Integration

AI border-control markets may involve several vertically connected levels:

Hardware

↓

Biometric sensors

↓

Identity database

↓

AI recognition model

↓

Risk-scoring system

↓

Cloud infrastructure

↓

Border-control platform

↓

Airlines / airports / governments

A company controlling multiple levels could use dominance at one level to foreclose rivals at another.

This creates potential vertical leveraging concerns.

21. Tying and Bundling

A dominant border technology provider could require:

"To purchase our biometric border gates, you must also use our AI analytics platform."

Or:

"Access to the passenger-information API is available only if the customer purchases our cloud-hosting service."

Such arrangements could raise tying or bundling issues where:

1. two separate products or markets exist;
2. the undertaking has substantial power in the tying product;
3. customers are effectively compelled to obtain the tied product;
4. the conduct has foreclosure potential; and
5. there is insufficient objective justification.

22. Procurement and Market Partitioning

Public procurement can also affect competition.

Suppose a government awards an exclusive nationwide contract covering:

• facial recognition;
• biometric databases;
• airport authentication;
• AI risk analytics; and
• cloud infrastructure.

A single procurement decision does not automatically constitute an antitrust violation.

But competition concerns may arise if:

• specifications unnecessarily favor one supplier;
• interoperability requirements exclude competitors;
• proprietary standards are mandated without justification;
• contract duration prevents market entry;
• switching is technically impossible.

23. Interoperability as a Competition Remedy

Possible remedies include:

1. API access

Require reasonable access to necessary interfaces.

2. Data portability

Permit lawful transfer of relevant information.

3. Open technical standards

Prevent unnecessary dependence on proprietary formats.

4. Non-discrimination

Require equivalent access conditions for competing providers.

5. Separation of functions

Separate infrastructure operation from downstream AI services.

6. Auditability

Permit regulatory examination of discriminatory algorithmic treatment.

7. Switching mechanisms

Ensure customers can migrate to alternative providers.

24. Objective Justification and National Security

Border control is fundamentally different from ordinary commercial digital platforms because national security and public safety may be directly implicated.

Consequently, an undertaking or public authority may argue that restrictions are necessary because of:

• terrorism prevention;
• identity fraud;
• cybersecurity;
• classified information;
• immigration enforcement;
• national-security legislation;
• protection of biometric information.

Competition law should therefore distinguish between:

genuine security restrictions

and

commercial restrictions presented as security restrictions.

The latter may receive greater scrutiny if they unnecessarily exclude competitors.

25. Important Analytical Test

A competition-law investigation could proceed through the following sequence:

Step 1 — Define the relevant market

Possible markets include:

• biometric border-control systems;
• automated passport control;
• border-risk analytics;
• identity verification;
• biometric databases;
• AI border-security software;
• airport security infrastructure.

Step 2 — Establish market power

Examine:

• market share;
• switching costs;
• network effects;
• data advantages;
• regulatory certification;
• interoperability;
• entry barriers.

Step 3 — Identify the partitioning mechanism

Determine whether partitioning results from:

• contract;
• algorithm;
• technical architecture;
• API restrictions;
• pricing;
• licensing;
• data access;
• procurement.

Step 4 — Assess foreclosure

Ask whether competitors are actually prevented or substantially impeded from competing.

Step 5 — Examine objective justification

Consider:

• national security;
• privacy;
• cybersecurity;
• regulatory compliance;
• technical necessity.

Step 6 — Proportionality

Even a legitimate security objective may not justify restrictions that go substantially beyond what is necessary.

26. Hypothetical Example

Assume Company X operates the dominant AI border-control platform used at 70% of a regional airport network.

Its platform controls:

• biometric authentication;
• passenger identity verification;
• AI risk scoring;
• API access to airport systems.

Company X launches its own AI analytics product.

It then:

1. gives its own analytics service real-time API access;
2. gives competitors delayed access;
3. charges competitors substantially higher access fees;
4. prevents data portability;
5. requires airports purchasing biometric gates to purchase its analytics product; and
6. uses territorial restrictions preventing its customers from switching to another provider in neighboring countries.

Potential theories could include:

• refusal of access;
• discriminatory access;
• tying;
• leveraging;
• interoperability foreclosure;
• exclusionary technical conduct;
• territorial market partitioning.

The strongest legal theory would depend on the relevant market, dominance, indispensability, actual effects and objective justifications.

27. Relationship Between the Major Cases

28. Key Legal Distinctions

Three distinctions are especially important.

Security restriction ≠ competition restriction

A border-control provider may legitimately restrict access to sensitive security information.

Proprietary technology ≠ essential facility

Ownership of advanced AI technology alone does not automatically create a duty to share it.

Market partitioning ≠ unlawful partitioning

Different national systems may legitimately exist because of different laws, security requirements and technical standards.

The competition issue becomes stronger where the partitioning is artificial, exclusionary and disproportionate, particularly when imposed by a dominant undertaking.

29. Emerging Competition Concerns

Future AI border-control systems may generate additional issues involving:

• AI model interoperability;
• biometric-data portability;
• federated identity systems;
• algorithmic discrimination;
• automated supplier selection;
• AI-generated exclusionary pricing;
• cross-border data silos;
• cloud concentration;
• sovereign AI infrastructure;
• cybersecurity certification;
• proprietary biometric standards;
• automated procurement;
• interoperability between national digital identities;
• AI model switching costs.

A particularly important issue is whether technical architecture itself becomes a competition instrument.

Traditional antitrust law often examines contracts and commercial behavior. AI infrastructure makes it increasingly necessary to examine the architecture through which market access is technically controlled.

30. Conclusion

AI border-control systems can become powerful digital gateways connecting governments, airports, airlines, identity providers, security companies and technology suppliers. Where one undertaking controls a critical gateway, competition concerns can arise if that control is used to partition markets, restrict interoperability, deny essential access, discriminate against rivals or extend dominance into adjacent AI services.

The central competition-law questions are:

1. Who controls the infrastructure?
2. Is the infrastructure genuinely indispensable?
3. Can competitors realistically replicate or bypass it?
4. Who controls the underlying data?
5. Are APIs and technical standards available on non-discriminatory terms?
6. Does the system favor affiliated services?
7. Are geographic restrictions legally necessary or commercially exclusionary?
8. Are security justifications genuine and proportionate?
9. Does the conduct foreclose effective competition?
10. Would interoperability, access or structural remedies restore competitive conditions?

The most directly useful precedents are Bronner, IMS Health, Commercial Solvents, Microsoft, Slovak Telekom, Google Shopping, Google Android, Magill, United Brands and Pierre Fabre. Collectively, they provide a framework for analysing how control over infrastructure, information, interoperability and territorial access can transform an AI border-control system from a technological tool into a potential mechanism of digital market partitioning.