Answer Bubbles: Information Exposure in AI-Mediated Search
Not on the map. Why?
Abstract (arXiv)
Generative search systems are increasingly replacing link-based retrieval with AI-generated summaries, yet little is known about how these systems differ in sources, language, and fidelity to cited material. We examine responses to 11,000 real search queries across five systems---vanilla GPT, Search GPT, Perplexity Search with Grok, Google AI Overviews, and traditional Google Search---at three levels: source diversity, linguistic characterization of the generated summary, and source-summary fidelity. We find that generative search systems exhibit significant textitsource-selection biases in their citations, favoring certain sources over others. Incorporating search also selectively attenuates epistemic markers, reducing hedging by up to 60% while preserving confidence language in the AI-generated summaries. At the same time, AI summaries further compound the citation biases: Wikipedia and longer sources are disproportionately overrepresented, whereas cited social media content and negatively framed sources are substantially underrepresented. Our findings highlight the potential for textitanswer bubbles, in which identical queries yield structurally different information realities across systems, with implications for user trust, source visibility, and the transparency of AI-mediated information access.
Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 21 of its 53 references that are in the map. First posted on arXiv 17 Mar 2026.
Builds on
21 of 53 references.
References · 21
- 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,266
- 2022BERTopic: Neural topic modeling with a class-based TF-IDF procedurecited by 779
- 2023FreshLLMs: Refreshing Large Language Models with Search Engine Augmentationcited by 308
- 2023Evaluating Verifiability in Generative Search Enginescited by 297
- 2018Dear Sir or Madam, May I introduce the GYAFC Dataset: Corpus, Benchmarks and Metrics for Formality Style Transfercited by 209
- 2025ZeroSearch: Incentivize the Search Capability of LLMs without Searchingcited by 203
- 2024Bias and Unfairness in Information Retrieval Systems: New Challenges in the LLM Eracited by 97
- 2023Neural Retrievers are Biased Towards LLM-Generated Contentcited by 37
- 2023Don't lose the message while paraphrasing: A study on content preserving style transfercited by 25
- 2022TyDiP: A Dataset for Politeness Classification in Nine Typologically Diverse Languagescited by 21
- 2025Collapse of Dense Retrievers: Short, Early, and Literal Biases Outranking Factual Evidencecited by 16
- 2024Dwell in the Beginning: How Language Models Embed Long Documents for Dense Retrievalcited by 13
9 more
- 2024Generative AI Search Engines as Arbiters of Public Knowledge: An Audit of Bias and Authoritycited by 12
- 2026MoltNet: Understanding Social Behavior of AI Agents in the Agent-Native MoltBookcited by 12
- 2025Auditing Google's AI Overviews and Featured Snippets: A Case Study on Baby Care and Pregnancycited by 7
- 2026Do We Know What They Know We Know? Calibrating Student Trust in AI and Human Responses Through Mutual Theory of Mindcited by 6
- 2025The Language of Approval: Identifying the Drivers of Positive Feedback Onlinecited by 4
- 2026The Hidden Toll of Social Media News: Causal Effects on Psychosocial Wellbeingcited by 4
- 2026Does Reasoning Make Search More Fair? Comparing Fairness in Reasoning and Non-Reasoning Rerankerscited by 2
- 2024Transparency, Privacy, and Fairness in Recommender Systemscited by 1
- 2024PromptDSI: Prompt-based Rehearsal-free Continual Learning for Document Retrievalcited by 1
Built on by
Citing papers per year: 2026: 3.
3 papers.