GenAI Advertising: Risks of Personalizing Ads with LLMs

2024-09-23

arXiv.org

AI Summary

This study investigates the ethical implications of embedding personalized advertisements in LLM-based chatbots, focusing on user experience and trust. Method: We designed and deployed three comparative systems—ad-integrated (unlabeled), ad-free, and ad-integrated with explicit disclosure—and conducted a controlled A/B test incorporating subjective user ratings, behavioral logs, and expert annotation. Contribution/Results: Our empirical analysis reveals that unlabeled ads temporarily improve perceived response desirability but significantly erode system credibility; 87% of users failed to detect ads autonomously, and upon disclosure, 62% perceived them as manipulative. Users consistently preferred managing privacy through conversational interaction rather than interface settings. The findings highlight three critical risks: high ad invisibility, substantial trust degradation, and weak user agency. These results provide empirically grounded warnings and design guidelines for ethically sustainable commercialization of LLMs.

Technology Category

• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Machine Learning: Privacy
• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness

Application Category

• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Recent advances in large language models have enabled the creation of highly effective chatbots, which may serve as a platform for targeted advertising. This paper investigates the risks of personalizing advertising in chatbots to their users. We developed a chatbot that embeds personalized product advertisements within LLM responses, inspired by similar forays by AI companies. Our benchmarks show that ad injection impacted certain LLM attribute performance, particularly response desirability. We conducted a between-subjects experiment with 179 participants using chabots with no ads, unlabeled targeted ads, and labeled targeted ads. Results revealed that participants struggled to detect chatbot ads and unlabeled advertising chatbot responses were rated higher. Yet, once disclosed, participants found the use of ads embedded in LLM responses to be manipulative, less trustworthy, and intrusive. Participants tried changing their privacy settings via chat interface rather than the disclosure. Our findings highlight ethical issues with integrating advertising into chatbot responses

Problem

• Investigating personalized ad integration in LLM chatbots for monetization
• Evaluating user detection and preference for hidden advertisements
• Developing ad-serving models and datasets for adaptive advertising

Innovation

• Embedding personalized ads into chatbot responses
• Fine-tuning Phi-4-Ads model for advertising adaptation
• Creating advertising dataset for personalized LLM monetization