[Paper Review] AI-Driven Contextual Advertising: A Technology Report and Implication Analysis

Emil Häglund, Johanna Björklund|arXiv (Cornell University)|May 2, 2022

Ethics and Social Impacts of AI4 citations

TL;DR

This paper examines AI-driven contextual advertising as a privacy-preserving alternative to behavioral targeting, leveraging AI for deeper semantic analysis of media context to improve ad relevance and effectiveness. It identifies key context factors—applicability, affective tone, and content involvement—while warning of risks like algorithmic bias and manipulative ad delivery if systems lack transparency and oversight.

ABSTRACT

Programmatic advertising consists in automated auctioning of digital ad space. Every time a user requests a web page, placeholders on the page are populated with ads from the highest-bidding advertisers. The bids are typically based on information about the user, and to an increasing extent, on information about the surrounding media context. The growing interest in contextual advertising is in part a counterreaction to the current dependency on personal data, which is problematic from legal and ethical standpoints. The transition is further accelerated by developments in Artificial Intelligence (AI), which allow for a deeper semantic understanding of context and, by extension, more effective ad placement. In this article, we begin by identifying context factors that have been shown in previous research to positively influence how ads are received. We then continue to discuss applications of AI in contextual advertising, where it adds value by, e.g., extracting high-level information about media context and optimising bidding strategies. However, left unchecked, these new practices can lead to unfair ad delivery and manipulative use of context. We summarize these and other concerns for consumers, publishers and advertisers in an implication analysis.

Motivation & Objective

• To analyze the role of AI in advancing contextual advertising as a privacy-compliant alternative to behavioral targeting.
• To identify key context factors—applicability, affective tone, and content involvement—that influence ad effectiveness and viewer perception.
• To evaluate the implications of AI-driven automation in contextual advertising, including risks of bias, discrimination, and lack of transparency.
• To explore opportunities for advertisers to leverage contextual targeting for broader audience reach and reactive campaign strategies.
• To advocate for explainable AI and multidisciplinary solutions to mitigate risks in automated ad delivery systems.

Proposed method

• Systematic review of existing literature on contextual advertising to extract core context factors influencing ad effectiveness.
• Analysis of AI applications in contextual advertising, including NLP, multimodal analysis, and real-time bidding optimization.
• Examination of AI's role in detecting high-level context features such as topic, sentiment, visual complexity, and engagement levels.
• Evaluation of automated bidding strategies in programmatic ad auctions using contextual signals instead of personal data.
• Identification of risks associated with AI automation, including proxy discrimination and lack of interpretability in decision-making.
• Discussion of regulatory and self-regulatory frameworks (e.g., IAB, CAN) to guide ethical deployment of AI in contextual advertising.

Experimental results

Research questions

• RQ1 Which context factors—applicability, affective tone, and content involvement—most significantly influence consumer perception of contextual ads?
• RQ2 How can AI enhance the precision and effectiveness of contextual advertising beyond keyword matching?
• RQ3 What are the primary risks of AI-driven contextual advertising, particularly regarding fairness, transparency, and manipulation?
• RQ4 In what ways can contextual advertising reduce privacy risks compared to behavioral targeting, and what new risks does it introduce?
• RQ5 How can advertisers effectively adopt AI-driven contextual advertising while ensuring ethical use and avoiding demographic bias?

Key findings

• AI enables deeper semantic understanding of media context, allowing for more precise ad placement based on topic, sentiment, and engagement levels.
• Contextual advertising can improve key metrics such as click-through rate, brand perception, and purchase intent when aligned with relevant content.
• The shift from behavioral to contextual targeting reduces reliance on personal data, mitigating privacy risks associated with third-party tracking.
• AI-driven systems may inadvertently discriminate by targeting context features that correlate with demographic traits like age, gender, or race.
• Lack of interpretability in AI models increases the risk of manipulative ad delivery and makes it difficult to detect or correct biased outcomes.
• Explainable AI and self-regulatory initiatives (e.g., IAB, CAN) are essential to ensure transparency and accountability in automated ad systems.