Ethics, Governance, and the Future of Media

Introduction

Media planning and buying are no longer concerned only with reach, frequency, cost, and return on investment. Contemporary media systems collect behavioral data, infer interests, automate decisions, deliver messages through complex supply chains, and influence public discourse. Consequently, media effectiveness cannot be separated from questions of legitimacy: Was the audience informed? Was consent meaningful? Were some groups unfairly excluded? Was the inventory genuine? Can the advertiser explain how a decision was made?

Ethical media management is therefore a strategic capability rather than a compliance exercise. A campaign that achieves short-term performance through intrusive surveillance, discriminatory targeting, misinformation, child exploitation, or opaque buying practices may create legal exposure, reputational damage, platform sanctions, and long-term loss of trust. This chapter examines the governance of media from the perspective of advertisers, agencies, publishers, platforms, regulators, and audiences. It also evaluates privacy-enhancing technologies, the decline of third-party cookies, agentic artificial intelligence, commerce media, sustainability, and future industry capabilities.

Learning Objectives

By the end of this chapter, readers should be able to:

1. Evaluate privacy, consent, discrimination, child protection, political advertising, and misinformation risks in media planning.
2. Explain ad fraud, invalid traffic, brand safety, transparency, and accountability across the media supply chain.
3. Compare regulatory and self-regulatory approaches to responsible advertising.
4. Assess privacy-enhancing technologies and strategic implications of cookie deprecation.
5. Design governance controls for AI-enabled, commerce-media, and programmatic activity.
6. Recommend measures that improve both campaign performance and social legitimacy.

1. From Media Performance to Media Legitimacy

A conventional media plan optimizes outcomes such as impressions, reach, conversions, or incremental sales. A governance-oriented plan adds constraints and conditions. It asks whether impressions are viewable, whether users are real, whether targeting is lawful and fair, whether content is suitable, and whether the environmental and social costs are acceptable.

A useful decision principle is:

Media value is not merely the value generated for the advertiser; it is the value generated without unjustified harm to audiences, publishers, competitors, or society.

This principle creates four dimensions of responsible media:

• Legality: compliance with applicable privacy, consumer-protection, competition, electoral, accessibility, and advertising rules.
• Ethical legitimacy: respect for autonomy, fairness, dignity, non-maleficence, and public accountability.
• Operational integrity: reliable measurement, authentic inventory, transparent transactions, and effective controls.
• Strategic sustainability: the capacity to maintain trust, performance, and societal acceptance over time.

Law establishes minimum requirements, but legal compliance does not automatically make a practice ethical. A consent banner may be technically lawful yet manipulative if it makes acceptance prominent and refusal difficult. Similarly, an algorithm may use permitted data but still reproduce discriminatory outcomes.

2. Privacy and Consent

2.1 The data lifecycle

Advertising data commonly passes through a lifecycle: collection, enrichment, segmentation, activation, measurement, retention, sharing, and deletion. Risk can arise at every stage. Data may be collected directly from a person, inferred from behavior, obtained from a partner, or generated by a device. The strategic question is not simply whether data are available, but whether the organization has a legitimate and clearly communicated basis for using them.

Important privacy principles include:

• Purpose limitation: collect and use data for specified purposes rather than indefinite future use.
• Data minimization: use only data necessary for the intended objective.
• Transparency: explain what is collected, why it is used, who receives it, and how long it is retained.
• Accuracy: provide mechanisms to correct materially inaccurate information.
• Security: protect data against unauthorized access, loss, and misuse.
• Retention limitation: delete or anonymize data when they are no longer needed.
• User rights: support access, correction, deletion, objection, portability, or restriction where applicable.

2.2 Consent as a governance process

Consent should be informed, specific, freely given, and capable of being withdrawn. A preselected box, confusing interface, bundled purposes, or refusal path that is materially harder than acceptance weakens consent quality. Consent is also not transferable without limits: permission given to one organization for one purpose does not automatically authorize unrelated uses by an entire advertising ecosystem.

Organizations should distinguish consent from other possible legal bases, such as contractual necessity, legal obligation, or legitimate interest. The appropriate basis depends on jurisdiction, context, data type, and balancing requirements. Sensitive data, precise location, biometric information, health information, and data about children generally require heightened safeguards.

2.3 Identity resolution and data matching

Identity graphs can connect devices, browsers, customer records, and household attributes. They can improve frequency management and measurement, but probabilistic matching may incorrectly merge individuals, while deterministic matching may create extensive profiles. Governance should specify acceptable match rates, error tolerance, permitted attributes, retention, and methods for honoring deletion or opt-out requests.

3. Fairness, Discrimination, and Inclusion

Targeting can produce discrimination even when an advertiser does not explicitly select a protected characteristic. Proxy variables such as postcode, language, device type, education, employment pattern, or browsing behavior may correlate with race, gender, age, disability, religion, or socioeconomic status.

Discrimination may occur through:

1. Targeting exclusion: preventing certain groups from seeing opportunities such as housing, employment, education, or credit.
2. Delivery optimization: an algorithm finding cheaper or more responsive audiences that disproportionately exclude protected groups.
3. Creative representation: portraying groups through stereotypes or failing to represent them.
4. Measurement bias: attributing conversions more accurately for some groups than others.
5. Pricing discrimination: charging different users or advertisers unfairly because of inferred characteristics.

A practical fairness review compares eligibility, reach, frequency, cost, delivery, and outcomes across relevant groups. A simple reach-rate diagnostic is:

where R_g is the reach rate for group g, I_g is the number of reached individuals in that group, and P_g is the eligible population of that group. The measure is descriptive, not proof of discrimination. It requires reliable population definitions, privacy-preserving aggregation, and attention to statistical uncertainty. Similar reach rates also do not prove fairness if creative quality, frequency, conversion opportunity, or exclusion rules differ.

Fairness is therefore a governance judgment involving law, social context, and stakeholder participation. Teams should document prohibited targeting categories, test delivery outcomes, involve legal and diversity specialists, and provide escalation procedures when automated delivery behaves unexpectedly.

4. Children and Vulnerable Audiences

Children may have limited ability to understand persuasive intent, data practices, or commercial incentives. Special risks include behavioral profiling, persuasive design, influencer marketing that is not clearly disclosed, collection of precise location, and advertising for age-restricted products.

Responsible controls include age-appropriate design, contextual rather than behavioral targeting, prohibition of sensitive profiling, clear separation between content and advertising, parental information where required, restricted frequency, and human review of creative. Age assurance itself creates privacy risks; collecting excessive identity evidence to verify age may be disproportionate. The preferred solution is usually the least intrusive method that can achieve the required protection.

Vulnerable adults may also require safeguards, particularly where advertising concerns gambling, high-cost credit, health treatments, political persuasion, or essential services. Vulnerability is contextual, not merely a fixed demographic category.

5. Misinformation, Disinformation, and Brand Safety

Misinformation is false or misleading information shared without necessarily intending harm. Disinformation is deliberately deceptive information. Advertising can fund harmful content indirectly when automated systems purchase inventory wherever audiences are available. It can also amplify false claims when advertisers use sensational environments for cheap reach.

Brand safety concerns adjacency to unsuitable content; brand suitability concerns whether an environment is appropriate for a particular brand, audience, and campaign objective. Controls include keyword and semantic classification, allowlists, blocklists, independent verification, inclusion and exclusion categories, supply-path restrictions, and rapid incident response. No classifier is perfect: keyword blocking can exclude legitimate journalism, while semantic systems can miss coded or emerging harmful content.

Advertisers should distinguish content risk from publisher viewpoint. Excessive blocking can reduce media diversity and disproportionately harm smaller publishers. Governance should therefore use documented risk tiers, proportional controls, and human review for borderline cases.

6. Ad Fraud and Invalid Traffic

Advertising fraud seeks to generate payment or value through deception. Invalid traffic (IVT) is traffic that does not represent legitimate advertising exposure or engagement. It includes bots, manipulated devices, hidden ads, stacked impressions, domain spoofing, click farms, malware-driven activity, and some forms of accidental or non-human traffic. Not all IVT is equally fraudulent; some may result from technical errors rather than deliberate criminal conduct.

Fraud risk is distributed across the supply chain. Programmatic auctions may involve advertisers, agencies, demand-side platforms, supply-side platforms, exchanges, resellers, verification firms, publishers, and connected devices. Each additional intermediary can increase opacity and fees.

Controls include:

• purchasing through authorized seller declarations and transparent supply paths;
• using ads.txt, app-ads.txt, and sellers.json where applicable;
• independent pre-bid and post-bid verification;
• frequency, geography, device, and placement anomaly monitoring;
• contractually defined invalid-traffic credits and audit rights;
• server-side log analysis and reconciliation of impression counts;
• rejecting implausibly cheap inventory and unexplained traffic spikes.

A basic financial diagnostic is the invalid-traffic rate:

The numerator and denominator must come from the same measurement scope and time period. Different vendors use different detection methodologies, so reported rates are not automatically comparable. A low detected rate does not prove that fraud is absent; it may indicate limited visibility or a narrow definition.

7. Transparency, Accountability, and Governance

Transparency means that relevant stakeholders can understand material aspects of a decision or transaction. It does not require revealing trade secrets or exposing personal data. In media buying, transparency may include the identity of sellers, fees, auction mechanics, data sources, targeting logic, verification methods, optimization rules, and measurement limitations.

Accountability goes further: an organization must be able to demonstrate that it identified risks, assigned responsibility, implemented controls, monitored outcomes, and corrected failures. A governance system should include:

A media governance board may include marketing, legal, privacy, procurement, information security, analytics, brand safety, and social-impact representatives. Its purpose is not to approve every creative variation but to set risk appetite, review high-risk use cases, and ensure that commercial incentives do not override basic safeguards.

8. Regulation and Self-Regulation

The regulatory environment varies across jurisdictions, but major legal themes recur: data protection, consumer deception, unfair commercial practices, children’s privacy, competition, accessibility, intellectual property, election communication, and platform accountability. Privacy regimes commonly impose duties concerning lawful processing, transparency, rights, security, and breach response. Consumer-protection regimes address misleading claims, hidden sponsorship, fake reviews, and dark patterns. Electoral rules may require sponsor identification, spending records, targeting restrictions, or public ad libraries.

Self-regulatory codes can supplement law through standards for disclosure, responsible targeting, comparative claims, children’s advertising, and influencer conduct. Industry standards are valuable when they are specific, independently monitored, and supported by meaningful sanctions. They are weak when membership is voluntary, definitions are vague, or audits are controlled by the entities being assessed.

Multinational advertisers should apply a jurisdictional matrix rather than assuming that one global policy satisfies every market. The strictest rule is not always appropriate everywhere, but a global baseline of privacy, fairness, safety, and documentation can reduce fragmented control.

9. Political Advertising and Democratic Risk

Political advertising differs from ordinary commercial advertising because it can influence elections, public policy, and civic trust. Risks include undisclosed sponsorship, microtargeting based on sensitive attributes, foreign interference, synthetic media, voter suppression, and rapid dissemination of false claims.

Responsible political-media governance should address sponsor verification, public ad archives, expenditure records, targeting restrictions, disclosure of synthetic or altered content, provenance, rapid correction procedures, and clear separation between civic information and persuasion. Platforms and agencies should not assume that automated moderation can resolve political truth disputes. Governance must combine law, independent oversight, procedural fairness, and transparency.

10. Sustainability and the Environmental Footprint of Media

Digital advertising consumes energy through data centers, content delivery, auctions, identity resolution, measurement, and device use. Repeated auctions and excessive supply-chain duplication can create emissions without proportionate audience value. Sustainability is therefore connected to media quality and efficiency.

A practical sustainability plan can reduce unnecessary bid requests, limit excessive creative weight, use efficient formats, consolidate supply paths, select lower-carbon infrastructure where credible evidence exists, and include environmental criteria in agency and publisher procurement. Emissions estimates should state boundaries and assumptions. For example, a campaign may measure media-related emissions as:

where each term represents estimated emissions from data processing, content delivery, audience-device activity, and operational activities. This is an accounting framework, not a universal physical law. Results depend on system boundaries, geographic energy mixes, allocation methods, and data quality. Claims such as “carbon neutral” require especially careful substantiation because offsets do not eliminate underlying emissions.

11. Privacy-Enhancing Technologies and Cookie Deprecation

Third-party cookies historically supported cross-site recognition, frequency management, retargeting, and attribution. Their decline reflects browser restrictions, regulation, consumer expectations, and platform strategy. Cookie deprecation does not end measurement or personalization; it shifts the industry toward first-party relationships, contextual signals, authenticated environments, cohort or modeled approaches, and privacy-enhancing technologies (PETs).

Important PETs include:

• Data clean rooms: controlled environments where parties compare or analyze data without freely exchanging raw records.
• Aggregation: reporting results in groups rather than exposing individual-level events.
• Differential privacy: adding calibrated statistical noise to reduce the chance that an individual can be identified from published results.
• Encryption and secure computation: protecting data during storage, transfer, or analysis.
• Federated learning: training models across decentralized data while limiting central collection of raw data.
• Contextual targeting: selecting content or environments based on page, program, or moment rather than a person’s cross-site history.
• Conversion modeling: estimating outcomes when direct observation is incomplete; models must disclose uncertainty and validation limitations.

PETs reduce risk but do not magically make a use ethical. A clean room can still support unfair exclusion, and aggregation can still reveal sensitive patterns when groups are small. Governance must cover purpose, access, re-identification risk, validation, retention, and accountability.

Cookie deprecation also changes strategic measurement. Advertisers should establish durable first-party value exchanges, consented customer data, media mix or incrementality testing, contextual planning, and carefully governed experiments. Overreliance on platform-reported attribution creates conflict of interest because the seller may measure the value of its own inventory.

12. Artificial Intelligence and Agentic Media Buying

AI already supports audience modeling, bid optimization, creative generation, fraud detection, forecasting, and customer service. Agentic AI refers to systems that can pursue goals through multiple steps, use tools, retrieve information, make decisions, and sometimes execute transactions with limited direct supervision.

An agentic media buyer might interpret a brief, select channels, create variants, request bids, adjust budgets, monitor results, and pause placements. This can improve speed and scale, but it creates new control risks: unauthorized spending, prompt injection, hidden optimization, hallucinated evidence, discriminatory delivery, leakage of confidential data, and unclear responsibility when several vendors’ agents interact.

A responsible AI control framework includes:

1. Defined authority: spending limits, approved channels, prohibited actions, and human approval points.
2. Traceability: logs of prompts, data, model versions, decisions, overrides, and outcomes.
3. Testing: performance, bias, robustness, privacy, security, and adversarial testing before deployment.
4. Human accountability: named owners who can intervene, explain, suspend, and remedy.
5. Vendor controls: audit rights, data-use restrictions, incident reporting, and model-change notification.
6. Continuous monitoring: drift detection, anomalous spend alerts, brand-safety review, and post-campaign audits.

The proper question is not whether AI is autonomous, but whether its autonomy is bounded, observable, reversible, and proportionate to the risk of its decisions.

13. Commerce Media and Retail Data Governance

Commerce media enables retailers, marketplaces, payment providers, and commerce platforms to monetize first-party purchase and browsing data. Its strategic strengths include proximity to transactions, closed-loop measurement, and high purchase relevance. Its risks include concentration of market power, conflicts between retailer interests and advertiser interests, opaque audience definitions, excessive data combination, and self-preferencing.

Advertisers should ask: What exactly is being measured? Is an attributed sale incremental? Are audiences deduplicated? Can the retailer alter ranking or inventory in ways that disadvantage competitors? Are data used for advertising also used for credit, pricing, or unrelated decisions? Contracts should specify data purposes, access, retention, clean-room rules, audit rights, and independent measurement.

14. An Integrated Responsible Media Planning Framework

A graduate-level media plan can incorporate the following sequence:

Step 1: Define the objective and public-risk profile. Identify whether the campaign concerns health, finance, politics, children, employment, housing, or other sensitive domains.

Step 2: Map the data and supply chain. Document data sources, processors, platforms, publishers, resellers, verification firms, and measurement dependencies.

Step 3: Classify audiences and decisions. Determine whether targeting, exclusion, personalization, or automated optimization could affect protected or vulnerable groups.

Step 4: Select the least intrusive effective method. Compare contextual, first-party, aggregated, and individual-level approaches.

Step 5: Establish controls and thresholds. Specify consent, brand-safety, IVT, reach, frequency, emissions, and escalation requirements.

Step 6: Validate before launch. Review creative claims, disclosures, pixels, data permissions, model behavior, seller authorization, and measurement design.

Step 7: Monitor during delivery. Inspect anomalies, unsuitable adjacency, demographic disparities, spend changes, invalid traffic, and complaints.

Step 8: Evaluate incrementality and harm. Separate observed attribution from causal evidence and assess unintended effects.

Step 9: Document and remedy. Preserve evidence, notify affected parties where required, correct errors, and update controls.

This framework changes optimization from “maximize conversions subject to budget” to “maximize legitimate value subject to financial, legal, ethical, environmental, and operational constraints.”

Applied Case: The Meridian Health Campaign

Background

Meridian, a fictional private healthcare provider, planned a digital campaign for a preventive screening service. Its agency proposed combining website visitors, third-party inferred health interests, location data, lookalike modeling, and automated retargeting. The initial plan promised efficient acquisition and strong attributed conversions.

Problem

Privacy review identified that inferred health interests could constitute sensitive data in some jurisdictions. The audience model also delivered disproportionately to affluent urban neighborhoods, while excluding older users who converted through offline channels. Ads appeared beside unreliable medical claims, and the platform’s attribution window overstated the campaign’s contribution because many users had already searched for the service.

Analysis

The campaign had four governance failures: excessive data use, possible sensitive profiling, unfair delivery, and weak causal measurement. It also created brand-safety and public-health risks. The agency’s low cost per acquisition concealed both exclusion and non-incremental conversions.

Approach and solution

Meridian replaced third-party health-interest segments with contextual placement and consented first-party audiences. It used broad geographic delivery with frequency limits, accessible creative, and offline booking options. A clean room supported aggregated overlap analysis, while an independent holdout test estimated incremental bookings. Medical content was reviewed by qualified professionals, and inventory was restricted to verified health and news environments. The contract added audit rights, deletion obligations, incident reporting, and a requirement to explain automated optimization.

Outcome

The revised plan produced a higher reported cost per attributed conversion but a more credible estimate of incremental bookings, broader reach among older audiences, fewer unsuitable placements, and lower regulatory exposure. Meridian also established a standing review committee for sensitive campaigns.

Learning points

1. Efficient targeting can be socially and strategically inefficient when it excludes valuable audiences.
2. Attribution is not the same as causation.
3. Sensitive categories require contextual and first-party alternatives.
4. Governance controls should be built before launch rather than added after a crisis.

Challenges and Limitations

No governance framework eliminates uncertainty. Consent may be legally valid but poorly understood. Fairness metrics can conflict, and protected-group data may be unavailable for legitimate privacy reasons. Fraud detection vendors disagree. Environmental estimates vary by boundary. AI systems change rapidly, and regulation often lags technical capability. Strict controls may reduce short-term performance or disadvantage smaller publishers.

Managers must therefore use proportionality and documented judgment. They should prioritize high-risk use cases, test alternatives, disclose uncertainty, and treat complaints and near misses as learning evidence. Ethical governance is not a one-time certification; it is an organizational capability supported by incentives, expertise, and leadership.

Chapter Summary

Responsible media planning extends beyond reach and return to include legitimacy, fairness, privacy, safety, sustainability, and accountability. Meaningful consent requires transparency, choice, purpose limitation, and withdrawal mechanisms. Targeting and automated delivery can create indirect discrimination, particularly in sensitive sectors. Children and vulnerable audiences require stronger safeguards. Misinformation, brand-safety failures, ad fraud, and invalid traffic damage both society and media efficiency, making verification and supply-chain transparency strategically important.

Regulation establishes minimum duties, while self-regulation and organizational governance can add operational standards. Political advertising requires heightened disclosure and democratic safeguards. Privacy-enhancing technologies, first-party data, contextual methods, clean rooms, and modeled measurement are increasingly important as third-party cookies decline. Agentic AI can improve scale but demands bounded authority, traceability, testing, and human accountability. Commerce media offers powerful transaction data but raises concentration and conflict-of-interest concerns. The future media professional must combine marketing performance with evidence-based risk management and public responsibility.

Key Terms

Privacy: Protection of information and decisions concerning identifiable or identifiable individuals.

Consent: A freely given, informed, specific, and withdrawable indication of agreement to data processing.

Purpose limitation: Restricting data use to specified and legitimate purposes.

Data minimization: Limiting collection and use to what is necessary.

Identity graph: A system linking identifiers across devices, channels, or records.

Discrimination: Unjustified differential treatment or impact affecting a person or group.

Brand safety: Protection from adjacency to unsuitable or harmful content.

Brand suitability: Assessment of whether an environment fits a particular brand and campaign.

Misinformation: False or misleading information not necessarily intended to deceive.

Disinformation: Deliberately deceptive or manipulative information.

Ad fraud: Deceptive activity intended to obtain advertising value or payment.

Invalid traffic: Traffic that does not represent legitimate advertising exposure or engagement.

Transparency: Availability of material information about decisions, transactions, and practices.

Accountability: Responsibility to implement controls, provide evidence, and remedy failures.

Privacy-enhancing technology: A technology that reduces privacy risk while enabling useful analysis or service.

Differential privacy: A method that limits individual disclosure risk through controlled statistical noise.

Clean room: A controlled environment for privacy-preserving data analysis or matching.

Agentic AI: AI capable of pursuing goals through multistep actions and tool use.

Commerce media: Advertising using data and inventory connected to commerce or transactions.

Incrementality: The causal outcome attributable to an intervention beyond what would otherwise occur.

Review Questions

Short Answer Questions

1. Why is legal compliance not always equivalent to ethical media practice?
2. Distinguish brand safety from brand suitability.
3. What is invalid traffic, and why is its measurement difficult?
4. Identify two privacy-enhancing technologies and describe their purpose.
5. Why does agentic AI require bounded authority and audit logs?

Long Answer Questions

1. Explain how a media campaign can create discrimination without explicitly using protected characteristics.
2. Evaluate the governance requirements for collecting and activating first-party data after third-party cookie deprecation.
3. Discuss the relationship among transparency, accountability, and supply-chain complexity in programmatic buying.
4. Compare the strategic benefits and ethical risks of commerce media.
5. Design a responsible media governance framework for a multinational advertiser operating across different regulatory environments.

Analytical / Application Questions

1. A financial-services campaign delivers lower costs when it excludes several postcodes. Develop an analysis to determine whether the optimization creates discriminatory access and recommend controls.
2. An advertiser’s platform report shows excellent attributed sales, but an independent holdout test shows little incremental lift. Explain the discrepancy and redesign the measurement approach.
3. An autonomous media agent begins reallocating budget toward sensational news environments and exceeds its approved spend. Identify the governance failures and specify technical, contractual, and managerial remedies.

Multiple Choice Questions with Answers

Question 1

Which principle requires an organization to collect only the data necessary for a stated advertising purpose?

A. Data minimization

B. Frequency capping

C. Attribution

D. Viewability

Correct Answer: A. Data minimization

Question 2

Which situation best illustrates indirect discrimination in media delivery?

A. Using a larger font in an advertisement

B. Using postcode optimization that systematically excludes a protected group

C. Showing the same ad twice to one user

D. Buying inventory from an authorized publisher

Correct Answer: B. Using postcode optimization that systematically excludes a protected group

Question 3

What is the principal difference between brand safety and brand suitability?

A. Safety concerns price, while suitability concerns reach

B. Safety concerns legal ownership, while suitability concerns frequency

C. Safety concerns harmful adjacency generally, while suitability concerns fit for a specific brand and campaign

D. There is no meaningful difference

Correct Answer: C. Safety concerns harmful adjacency generally, while suitability concerns fit for a specific brand and campaign

Question 4

Which measure is most directly intended to identify non-human or illegitimate advertising activity?

A. Invalid-traffic rate

B. Gross rating point

C. Share of voice

D. Net promoter score

Correct Answer: A. Invalid-traffic rate

Question 5

Which technology enables parties to analyze overlapping data while restricting direct exchange of raw records?

A. A clean room

B. A billboard

C. A content delivery network

D. A frequency cap

Correct Answer: A. A clean room

Question 6

Why is a platform-reported attributed conversion not necessarily an incremental conversion?

A. Attribution always excludes digital media

B. The user may have converted without the advertising exposure

C. Incrementality cannot be measured experimentally

D. Attributed conversions are always fraudulent

Correct Answer: B. The user may have converted without the advertising exposure

Question 7

Which control is most appropriate for an agentic media-buying system?

A. Unlimited autonomous spending

B. Removal of all activity logs

C. Defined authority, approval thresholds, and intervention capability

D. Allowing the model to change its own objectives without review

Correct Answer: C. Defined authority, approval thresholds, and intervention capability

Question 8

Which practice is generally most privacy-preserving for reaching users based on the content environment rather than cross-site identity?

A. Contextual targeting

B. Undisclosed health profiling

C. Unrestricted identity matching

D. Selling raw browsing histories

Correct Answer: A. Contextual targeting

Question 9

Why can a low detected IVT rate fail to prove that fraud is absent?

A. IVT is unrelated to measurement

B. Detection methods and visibility may be limited

C. All impressions are automatically valid

D. Fraud occurs only in offline media

Correct Answer: B. Detection methods and visibility may be limited

Question 10

What is a central governance concern in political advertising?

A. Using any sponsor disclosure possible

B. Undisclosed influence, sensitive microtargeting, and synthetic content

C. Reducing all political information to entertainment

D. Eliminating public ad archives

Correct Answer: B. Undisclosed influence, sensitive microtargeting, and synthetic content