Competition Law and Competition Issues in Cognitive Advertising Ecosystems

1. Introduction

Cognitive advertising ecosystems are advertising systems that use artificial intelligence, machine learning, behavioural data, psychographic profiling, natural-language processing, emotion or attention analysis, predictive analytics, recommendation systems, and automated decision-making to understand consumers and deliver highly personalised advertisements.

Traditional advertising generally asks:

“Which group of consumers should receive this advertisement?”

Cognitive advertising goes further and attempts to determine:

“What is this particular consumer likely to think, want, notice, click, or purchase, and what advertisement should be shown at this precise moment?”

This creates substantial commercial efficiencies, but it also creates important competition-law issues. The same data, algorithms, computing infrastructure and platforms that improve advertising efficiency can also create barriers to entry, strengthen dominant platforms, facilitate exclusion of competitors, and increase dependence on a small number of advertising ecosystems.

Competition authorities have increasingly examined the relationship between data, advertising technology, platforms, algorithms and market power. For example, the EU's Meta/Facebook litigation recognised that competition authorities may consider data-protection issues when assessing competition-law questions, while U.S. authorities have brought major proceedings concerning digital advertising technology. (Eur-Lex)

2. Meaning of Cognitive Advertising

Cognitive advertising may be understood as:

The use of data-driven computational systems to infer consumer interests, preferences, behaviour, attention or likely responses and to automatically select, personalise, price, place or optimise advertising.

It may use:

Artificial intelligence

Machine learning

Natural-language processing

Behavioural targeting

Predictive analytics

Consumer profiling

Sentiment analysis

Emotion recognition

Contextual advertising

Recommendation algorithms

Real-time bidding

Automated ad placement

Facial or biometric information in some applications

Cross-device tracking

Location and browsing data

Purchase histories

Social-media interactions

Search behaviour

Thus, cognitive advertising is not merely an advertising technique. It can become an entire economic ecosystem connecting:

Consumers → Data → AI models → Advertisers → Ad exchanges → Publishers → Measurement systems → Payment systems

3. What Is a Cognitive Advertising Ecosystem?

A cognitive advertising ecosystem normally contains several interconnected layers.

A. Consumer layer

Consumers generate data through:

searches;

clicks;

purchases;

videos watched;

websites visited;

social-media activity;

app usage;

location;

device information;

interactions with advertisements.

B. Data layer

The collected information is converted into:

consumer profiles;

audience segments;

behavioural predictions;

purchase probabilities;

interest scores;

attention scores.

C. AI/algorithmic layer

Algorithms determine:

which advertisement to show;

when to show it;

where to show it;

how much an advertiser should bid;

which consumers are valuable;

which advertisement is predicted to generate the greatest response.

D. Advertising-market layer

Advertisers, publishers, demand-side platforms, supply-side platforms and ad exchanges interact.

E. Measurement layer

The system evaluates:

impressions;

clicks;

conversions;

engagement;

customer acquisition;

return on advertising expenditure.

The feedback is then used to improve the algorithm.

This creates a self-reinforcing data-and-algorithmic cycle.

4. Why Cognitive Advertising Creates Competition Issues

The central competition problem is that data and intelligence can become sources of market power.

A company with:

enormous amounts of consumer data,

powerful AI models,

a large user base,

extensive advertising inventory,

an advertising exchange,

an ad server,

measurement tools,

may have advantages that smaller competitors cannot easily reproduce.

Therefore, competition authorities may need to examine not merely market share but also:

Who controls the data, algorithms, interfaces, advertising inventory and access points?

5. Major Competition Issues

5.1 Data concentration

Data is one of the most important inputs into cognitive advertising.

A dominant platform may collect data from:

its search engine;

social network;

video service;

mobile operating system;

browser;

applications;

advertising network;

third-party websites.

Combining these datasets can produce extremely detailed consumer profiles.

A rival may therefore face a substantial disadvantage because it cannot obtain comparable data.

Competition concern

Data concentration may create:

barriers to entry;

economies of scale;

economies of scope;

stronger targeting capabilities;

higher advertiser dependence;

reduced ability of rivals to compete.

The German Facebook proceeding is particularly important here. The Bundeskartellamt examined Facebook's combination of data from Facebook, affiliated services and third-party sources and treated the data practices as relevant to abuse of dominance. (Bundeskartellamt)

6. Algorithmic Market Power

Cognitive advertising depends heavily upon algorithms.

An advertising algorithm may determine:

consumer segmentation;

advertising ranking;

bid allocation;

price;

visibility;

recommendation;

conversion probability.

If a dominant platform controls the algorithm, it may indirectly control access to consumers.

The problem becomes more serious when the same company controls:

consumer data;

advertising inventory;

advertiser relationships;

advertising exchange;

ad server;

measurement tools;

algorithmic ranking.

This creates a potential vertical concentration of the advertising ecosystem.

7. Self-Preferencing

A platform may use its control over an advertising ecosystem to favour its own products.

For example, a platform could theoretically:

give its own advertising products better access;

favour its own ad exchange;

provide better data to its own advertising division;

rank its own advertisements or services more favourably;

restrict rival advertising technology.

This is particularly important where the platform operates simultaneously as:

competitor + infrastructure provider + intermediary + data controller.

The EU Google Shopping case demonstrates the broader competition-law principle that a dominant digital platform's control over ranking and visibility can raise Article 102 TFEU concerns when it systematically favours its own service over competing services. (Eur-Lex)

8. Exclusive Dealing and Foreclosure

Cognitive advertising platforms may attempt to secure exclusive relationships with:

publishers;

advertisers;

agencies;

app developers;

data suppliers;

websites.

Exclusive arrangements may prevent rivals from obtaining sufficient:

advertising inventory;

consumer data;

audience information;

publisher access;

transaction volume.

The result can be foreclosure of competing advertising platforms.

The Google AdSense litigation is particularly relevant because it concerned contractual restrictions and exclusive-supply obligations in online search advertising intermediation. In Case T-334/19, the General Court examined Google's contractual restrictions under Article 102 TFEU. (Eur-Lex)

9. Real-Time Advertising Auctions

Modern advertising often operates through automated auctions.

When a user opens a webpage, an automated system may conduct an auction in milliseconds.

A simplified process is:

User visits webpage → advertising opportunity created → data analysed → advertisers bid → algorithm evaluates bids → advertisement selected → advertisement displayed

This raises competition questions concerning:

auction neutrality;

bid manipulation;

preferential treatment;

access to information;

conflicts of interest;

discriminatory rules;

transparency;

algorithmic coordination.

The U.S. Department of Justice's Google ad-tech case specifically alleged anticompetitive conduct involving multiple parts of the digital advertising technology stack and auction mechanisms. (Justice.gov)

As of September 2026, the U.S. Department of Justice reported that the federal district court had ordered substantial behavioural relief concerning Google's advertising-technology markets, including measures intended to facilitate interoperability with rival solutions. (Justice.gov)

10. Algorithmic Collusion

Cognitive advertising systems may use sophisticated algorithms to determine:

advertising prices;

bids;

commissions;

inventory prices;

targeting prices.

A competition concern arises if algorithms facilitate coordinated behaviour between competitors.

Example

Suppose several advertising platforms independently use algorithms that continuously observe competitors' prices and rapidly adjust their own prices.

Even without an explicit human agreement, algorithms can potentially make coordination easier.

However, the mere use of similar algorithms is not automatically illegal. Competition law normally requires examination of the actual conduct, evidence of coordination and applicable legal standards.

The important principle is that technological automation does not itself provide immunity from competition law.

11. Personalisation as a Competitive Advantage

Cognitive advertising can generate powerful network effects.

More users create:

More data → better predictions → better advertising performance → more advertisers → more revenue → greater investment → better algorithms → more users

This is a feedback loop.

A smaller competitor may have:

Fewer users → less data → weaker predictions → fewer advertisers → less revenue → less investment

Consequently, the market can potentially become concentrated even without traditional exclusive dealing.

12. Data-Based Barriers to Entry

A new advertising platform may need:

millions of users;

extensive behavioural data;

sophisticated AI;

advertiser relationships;

publisher relationships;

large-scale computing;

measurement systems.

This can make entry expensive.

The important competition-law question becomes:

Is the advantage the result of legitimate innovation and economies of scale, or has a dominant undertaking used exclusionary conduct to prevent competitors from obtaining necessary scale?

That distinction is essential.

13. Network Effects

Cognitive advertising ecosystems often have multiple sides:

Consumer side

Consumers provide attention and data.

Advertiser side

Advertisers purchase advertising opportunities.

Publisher side

Publishers provide advertising inventory.

Data side

Data providers supply information.

Technology side

Ad-tech companies provide infrastructure.

These sides reinforce each other.

A large consumer base attracts advertisers.

More advertisers generate more revenue.

More revenue permits investment in technology.

Better technology can attract more users and publishers.

This can produce substantial network effects and economies of scale.

14. Multi-Sided Market Problems

Traditional competition analysis often focuses on one market.

Cognitive advertising ecosystems require examination of several interconnected sides.

For example:

Consumers ↔ Social platform ↔ Advertisers

or:

Users ↔ Search engine ↔ Advertisers

or:

Publishers ↔ Ad exchange ↔ Advertisers

The U.S. Supreme Court's decision in Ohio v. American Express Co., 585 U.S. 529 (2018) is important for understanding two-sided transaction platforms.

The case demonstrates that competition analysis of a multi-sided platform may need to consider the relationship between different sides of the platform rather than examining one side in isolation.

This is highly relevant to cognitive advertising because the platform simultaneously serves consumers, advertisers and other ecosystem participants.

15. Quality and Privacy as Competition Parameters

Competition is not always about price.

In digital advertising, consumers may receive services for little or no monetary price.

Instead, competition may occur through:

privacy;

security;

advertising intensity;

user experience;

data collection;

transparency;

personalisation.

This is important because an advertising platform may theoretically worsen privacy conditions without immediately increasing a monetary price.

The Facebook litigation illustrates this issue. The FTC's case described Facebook's business model as surveillance-based advertising and argued that diminished privacy and more intrusive advertising could be relevant to competitive conditions. (Federal Trade Commission)

16. Data Combination and Competition

A dominant company may possess data from several services.

For example:

Search data + social-media data + location data + purchase data + browsing data

can generate more detailed advertising predictions than any individual dataset.

The competition issue is therefore not simply:

“How much data does the company possess?”

but:

“What competitive advantage results from combining datasets that rivals cannot reasonably reproduce?”

The CJEU's decision in Meta Platforms v Bundeskartellamt, Case C-252/21 (2023) is particularly significant because it addressed the relationship between competition law and personal-data processing in the context of a dominant social-network operator. (Eur-Lex)

17. Consumer Lock-In

Cognitive advertising ecosystems can create switching costs.

Users may have:

extensive personal histories;

personalised recommendations;

social connections;

stored preferences;

customised advertising experiences;

linked applications.

Advertisers may similarly have:

historical campaign data;

audience segments;

conversion histories;

measurement systems;

optimisation models.

Moving to another platform may therefore be costly.

This can reduce multi-homing and strengthen incumbent market power.

18. Acquisitions of Potential Competitors

A dominant advertising ecosystem may acquire:

AI startups;

ad-tech companies;

data analytics companies;

identity-management businesses;

recommendation companies;

consumer-data platforms.

Competition law must examine whether such acquisitions eliminate future competitive threats.

The FTC's Facebook case is relevant because it alleged that Facebook's acquisitions of Instagram and WhatsApp were part of a strategy to eliminate competitive threats and maintain monopoly power. (Federal Trade Commission)

The lesson for cognitive advertising is broader:

An acquisition may be competitively significant even when the target is not a major advertising competitor at the moment of acquisition.

19. Tying and Bundling

A dominant ecosystem might bundle:

advertising services;

analytics;

identity services;

data management;

cloud services;

search;

social media;

ad exchange services.

For example, access to one service could be conditioned upon purchasing another service.

This can make it difficult for independent competitors to compete with individual components.

The Google Android litigation is useful by analogy. The General Court examined Google's use of agreements involving product bundling, exclusivity payments and anti-fragmentation obligations within a multi-sided ecosystem. (Eur-Lex)

20. Interoperability Restrictions

Advertising ecosystems depend on interoperability.

A dominant platform could potentially restrict:

API access;

data portability;

third-party measurement;

interoperability with rival ad exchanges;

access to advertising inventory.

This may make it difficult for competitors to enter or expand.

The Facebook antitrust litigation is relevant because the FTC alleged restrictions on API access that hindered competing applications and contributed to the maintenance of Facebook's position. (Federal Trade Commission)

21. Important Case Laws

Case 1: Meta Platforms v Bundeskartellamt

Case C-252/21, CJEU, 2023

Facts

The German competition authority examined Facebook's combination of user data from Facebook and other sources.

Legal issue

Whether data-processing practices of a dominant digital platform could be relevant to competition-law enforcement.

Importance

The CJEU addressed the relationship between competition law and GDPR-related questions. It confirmed that competition authorities may, under the circumstances of the case, take GDPR requirements into consideration when assessing abuse of dominance. (Eur-Lex)

Relevance to cognitive advertising

This case is highly relevant to:

behavioural advertising;

profiling;

cross-platform data;

personalised advertising;

data concentration;

privacy as a competition parameter.

Case 2: Google and Alphabet v European Commission

Case T-334/19 — Google AdSense for Search

Facts

The European Commission examined Google's contractual restrictions in online search advertising intermediation.

Legal issue

Whether contractual restrictions imposed by a dominant undertaking could restrict competition.

Importance

The General Court's 2024 judgment concerned Google's alleged abuse of dominance in online search advertising intermediation and examined exclusive-supply obligations and contractual restrictions. (Eur-Lex)

Relevance

The case demonstrates how advertising ecosystems can create competition concerns through:

exclusivity;

contractual restrictions;

control over advertising intermediation;

foreclosure of competing advertising services.

Case 3: United States v. Google LLC — Digital Advertising Technology

U.S. District Court for the Eastern District of Virginia

Facts

The U.S. Department of Justice and several states sued Google in 2023, alleging monopolisation of important digital advertising technology markets. (Justice.gov)

The government alleged that Google's position across different parts of the advertising technology stack allowed it to influence competition between publishers, advertisers and intermediaries.

Importance

The case directly demonstrates the competition significance of:

ad exchanges;

ad servers;

auction systems;

publisher technology;

advertiser technology;

vertical integration.

The court ordered significant relief in September 2026 concerning Google's advertising-technology practices, according to the DOJ. (Justice.gov)

Relevance

For cognitive advertising, the case illustrates why control of the technical infrastructure behind automated advertising decisions can become a competition-law issue.

Case 4: Google Search / Google Shopping

European Commission, 2017; General Court litigation

Facts

Google was found to have given more favourable positioning and display to its own comparison-shopping service than to competing comparison-shopping services.

Legal principle

A dominant platform controlling a major gateway can potentially affect competition by using that gateway to favour its own service.

The European Commission's decision identified preferential positioning and display of Google's comparison-shopping service as an infringement of Article 102 TFEU. (Eur-Lex)

Relevance to cognitive advertising

The principle can be relevant where an AI-driven advertising or recommendation platform:

controls visibility;

controls ranking;

competes with third-party services;

uses its algorithmic position to favour its own products.

Case 5: Ohio v. American Express Co.

585 U.S. 529 (2018)

Facts

American Express operated a two-sided transaction platform involving merchants and cardholders.

Legal principle

The Supreme Court examined the relevant market in the context of a two-sided platform.

Relevance

Cognitive advertising is similarly multi-sided.

A platform may connect:

consumers;

advertisers;

publishers;

data providers.

Therefore, the effects on one side may influence the economics of another side.

The case is especially useful for explaining why competition analysis cannot always focus exclusively on advertisers while ignoring consumers or publishers.

Case 6: FTC v. Facebook / Meta

U.S. District Court for the District of Columbia

Facts

The FTC alleged that Facebook maintained its personal social-networking monopoly through acquisitions and restrictive conduct involving developers.

The FTC specifically linked Facebook's market position to its advertising business and alleged that reduced competition affected advertisers as well as consumers. (Federal Trade Commission)

Legal relevance

The case demonstrates:

network effects;

acquisitions of potential competitors;

platform access;

API restrictions;

data-driven advertising;

barriers to entry.

Relevance to cognitive advertising

A platform that controls the social network may simultaneously control:

users + data + attention + advertising inventory + AI targeting

This can make competitive foreclosure particularly important.

Case 7: Google Android

Case T-604/18, General Court, 2022

Facts

The European Commission examined Google's conduct involving Android devices, Google Search, Chrome, the Play Store and agreements with device manufacturers and mobile network operators.

The General Court considered concepts including:

multi-sided platforms;

ecosystems;

product bundling;

exclusivity payments;

anti-fragmentation obligations;

exclusionary effects. (Eur-Lex)

Relevance

Cognitive advertising ecosystems increasingly operate across:

operating system + browser + search + applications + data + advertising.

Therefore, control over an ecosystem can have implications beyond a single advertising market.

Case 8: Microsoft Corp. Antitrust Litigation

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

Importance

The Microsoft litigation remains a foundational authority concerning:

platform power;

network effects;

exclusionary conduct;

interoperability;

control over technological gateways.

Relevance to cognitive advertising

AI advertising ecosystems can similarly depend upon technological gateways.

For example:

Operating system → browser → search → identity → data → advertising

If one firm controls several critical layers, rivals may face difficulties accessing users.

The case therefore provides a useful conceptual framework for analysing platform foreclosure.

22. Indian Competition-Law Perspective

Cognitive advertising can fall within the framework of the Competition Act, 2002, depending upon the facts.

Section 3

Section 3 addresses agreements that cause or are likely to cause an appreciable adverse effect on competition.

Potential issues include:

algorithmic coordination;

agreements among advertising platforms;

information sharing;

restrictive arrangements;

exclusive arrangements.

Section 4

Section 4 concerns abuse of dominant position.

Potential conduct includes:

discriminatory access;

unfair conditions;

denial of market access;

leveraging;

tying;

exclusionary conduct.

Sections 5 and 6

These provisions concern combinations and therefore become relevant to acquisitions involving:

AI advertising companies;

data companies;

ad-tech companies;

recommendation companies;

consumer analytics businesses.

23. Competition and Artificial Intelligence

AI changes advertising competition in several ways.

Traditional advertising

Human decision-making:

Advertiser → Agency → Advertisement → Consumer

Cognitive advertising

Automated decision-making:

Consumer data → AI model → Prediction → Auction → Advertisement → Feedback → Model improvement

The feedback mechanism is crucial.

The system continuously learns.

Consequently, a company with a larger data pool can potentially improve its model faster.

This can create a data-learning competitive advantage.

24. Algorithmic Transparency

Competition authorities may face difficulty determining whether an algorithm is:

neutral;

discriminatory;

self-preferencing;

exclusionary;

coordinated.

AI systems may be extremely complex.

This creates an enforcement problem:

How can competition authorities prove anti-competitive conduct when the relevant decision is made by a complex machine-learning system?

Possible evidence may include:

internal documents;

algorithmic outputs;

historical auction data;

A/B testing;

access rules;

pricing data;

internal communications;

model objectives;

technical audits.

25. Algorithmic Bias and Competitive Discrimination

An advertising algorithm might systematically favour certain advertisers.

Possible reasons could include:

higher bids;

strategic partnerships;

preferential contracts;

platform ownership;

data advantages;

algorithmic design.

Competition law may become relevant where discriminatory treatment excludes competitors rather than merely reflecting legitimate commercial differences.

The critical question is:

Does the algorithm reflect competition on the merits, or does it systematically disadvantage competing businesses?

26. Consumer Attention as a Scarce Resource

Cognitive advertising competes not only for consumer money but also for:

attention;

time;

engagement;

emotional response.

AI systems may optimise advertising for maximum attention.

This produces a new competitive resource:

Consumer attention

A platform controlling large quantities of consumer attention may have significant bargaining power over advertisers.

27. Advertising Quality Competition

Competition may occur through:

less intrusive advertising;

greater relevance;

better privacy;

better transparency;

fewer advertisements;

better consumer experience.

Therefore, competition authorities should not focus exclusively on advertising prices.

Quality is also a competitive parameter.

28. Exclusive Data Access

Suppose a dominant platform prevents advertisers or publishers from accessing data generated through the platform.

This may increase dependency.

Competitors may be unable to replicate:

audience segmentation;

conversion prediction;

consumer profiling;

campaign optimisation.

The competition question is whether restricting access is legitimate protection of privacy/security or whether it unnecessarily excludes competitors.

29. Interoperability and Data Portability

Competitive advertising markets may benefit from:

interoperable measurement systems;

portability of advertising data;

open APIs;

independent verification;

cross-platform measurement.

However, data portability must also respect:

privacy law;

cybersecurity;

intellectual-property rights;

confidentiality.

Competition law therefore interacts closely with other regulatory fields.

30. Merger-Control Issues

Cognitive advertising increases the importance of acquisitions involving small AI/data companies.

Traditional turnover thresholds may not fully capture the competitive significance of a startup holding:

unique consumer data;

advanced AI models;

behavioural datasets;

novel targeting technology.

Competition authorities may therefore need to examine:

potential competition;

innovation competition;

data concentration;

future market development;

elimination of emerging rivals.

31. Leveraging

A dominant company in one market might leverage its position into advertising.

For example:

Dominant search engine → advertising data → advertising exchange

or:

Dominant operating system → user data → advertising service

or:

Dominant social network → consumer attention → advertising technology

This raises the possibility of extending market power from one ecosystem layer into another.

32. Foreclosure of Smaller Advertisers and Ad-Tech Firms

Small firms may depend upon dominant platforms for access to customers.

If the platform changes:

ranking;

commission;

API rules;

auction rules;

data access;

targeting requirements,

smaller firms may suffer substantial competitive disadvantages.

This creates a potential platform dependency problem.

33. Competition Between AI Advertising Models

Future competition may occur between:

large foundation models;

specialised advertising models;

open-source AI;

proprietary recommendation engines;

consumer-side AI assistants;

autonomous advertising agents.

If one ecosystem controls the dominant AI model and the dominant advertising inventory, it may have advantages in both:

prediction capability + distribution capability.

That combination may become a significant source of market power.

34. Cognitive Advertising and Consumer Manipulation

Consumer manipulation is primarily a consumer-protection issue, but it can also have competition implications.

For example, highly personalised systems could make it difficult for consumers to:

compare products;

discover alternative suppliers;

switch platforms;

understand commercial recommendations.

Where such conduct is linked to exclusionary behaviour by a dominant firm, competition law may become relevant.

35. Privacy as a Competitive Variable

Privacy can function as a dimension of competition.

Two platforms may offer similar services but differ in:

amount of tracking;

data retention;

cross-platform profiling;

advertising personalisation.

If consumers value privacy, deterioration of privacy may constitute a reduction in non-price competition.

The Meta/Facebook proceedings demonstrate why the relationship between privacy, data processing and competition is increasingly important in digital markets. (Bundeskartellamt)

36. Key Legal Tests

Competition authorities should generally examine:

1. Relevant market

What market is affected?

2. Market power

Does the undertaking possess substantial market power?

3. Data advantage

Does it control unique or difficult-to-replicate data?

4. Network effects

Does the ecosystem benefit from strong network effects?

5. Conduct

What exactly has the undertaking done?

6. Foreclosure

Does the conduct restrict rivals' ability to compete?

7. Consumer effects

Are price, quality, privacy, innovation or choice affected?

8. Efficiencies

Does the conduct produce legitimate efficiencies?

9. Less restrictive alternatives

Could the same efficiency be achieved through less exclusionary means?

37. Pro-Competitive Benefits of Cognitive Advertising

Competition law should not treat cognitive advertising as inherently anti-competitive.

It may produce significant benefits.

Benefits include:

Better matching between consumers and products.

Lower advertising waste.

Improved advertiser targeting.

Better measurement.

Lower customer-acquisition costs.

Improved advertising relevance.

Greater opportunities for small advertisers.

Better publisher monetisation.

Faster experimentation.

Innovation in marketing technology.

The competition issue is therefore not AI advertising itself, but whether market power is acquired or maintained through exclusionary or exploitative conduct.

38. Difference Between Innovation and Anti-Competitive Conduct

A company may legitimately develop a superior AI advertising model.

That is generally competition on the merits.

The concern becomes stronger where the dominant company combines innovation with conduct such as:

exclusionary agreements;

discriminatory access;

self-preferencing;

tying;

foreclosure;

acquisition of nascent competitors;

manipulation of advertising auctions;

restriction of interoperability.

Therefore:

Better algorithm ≠ automatically anti-competitive conduct.

But:

Better algorithm + exclusionary conduct + market power = potentially serious competition-law issue.

39. Regulatory Challenges

Competition authorities face several challenges.

A. Defining the market

Advertising ecosystems contain several interconnected markets.

B. Measuring data advantages

Traditional market shares may not capture data-based power.

C. Understanding AI

Authorities require technical expertise.

D. Establishing causation

It may be difficult to prove that an algorithm caused competitive foreclosure.

E. Distinguishing efficiency from exclusion

A sophisticated algorithm may legitimately improve performance.

F. Rapid technological change

AI markets may change faster than traditional litigation.

40. Overall Competition-Law Framework

The competition analysis of cognitive advertising can be represented as:

Consumer Data

↓

Profiling and Prediction

↓

AI/Algorithmic Decision-Making

↓

Advertising Auction

↓

Advertisement Placement

↓

Consumer Response

↓

New Data

↓

Improved AI Model

This creates a continuous data-feedback loop.

If a dominant company controls the entire loop, competitors may find it difficult to obtain sufficient scale.

41. Short Case-Law Revision Table

42. Important Exam Points

For examination purposes, remember these 10 major competition issues:

Data concentration

Algorithmic market power

Self-preferencing

Exclusive dealing

Advertising-auction manipulation

Algorithmic coordination

Network effects

Barriers to entry

Interoperability restrictions

Acquisition of potential AI/data competitors

43. Conclusion

Cognitive advertising ecosystems represent a major evolution from traditional advertising because they combine consumer data, artificial intelligence, behavioural prediction, automated auctions and platform infrastructure.

Their competition-law significance arises from the possibility that control over data and AI systems can become a source of durable market power.

The central legal questions are therefore:

Who controls the consumer data?

Who controls the advertising inventory?

Who controls the algorithm?

Who controls the auction?

Can rival platforms obtain comparable data?

Can advertisers and publishers switch easily?

Is access to the ecosystem discriminatory?

Are competitors being foreclosed?

Are acquisitions eliminating future competitors?

Are privacy and quality being reduced as dimensions of competition?

The modern approach is consequently moving beyond simple price-and-market-share analysis toward examination of data, algorithms, network effects, interoperability, ecosystem control and non-price competition.

The major cases involving Meta/Facebook, Google AdSense, Google Shopping, Google AdTech, Google Android, American Express and Microsoft provide useful legal principles for analysing these issues. In particular, the Meta and Google proceedings demonstrate how data and advertising infrastructure can become central competition-law concerns in digital ecosystems. (Eur-Lex)