How Generative AI Disrupts Search: An Empirical Study of Google Search, Gemini, and AI Overviews

A comprehensive empirical study reveals that generative AI is fundamentally reshaping web search by retrieving different sources and presenting information differently than traditional search engines. The research finds that AI Overviews appear in over half of queries, tend to prioritize Google-owned content over institutional sources, and show lower consistency and robustness compared to standard search results.

This research exposes a critical tension in the integration of generative AI into search ecosystems. While AI Overviews provide convenience by synthesizing information directly for users, the study reveals they operate under different algorithmic priorities than traditional search engines. The 51.5% appearance rate for AI Overviews on representative queries indicates this is not a marginal feature but a core infrastructure shift. The disparity in source selection—particularly the preference for Google-owned content over government and educational institutions—suggests potential conflicts of interest that could distort information access.

The finding that websites blocking AI crawlers are systematically disadvantaged despite Google having access to their content raises profound questions about digital fairness and publisher control. This creates a perverse incentive structure where publishers must choose between participating in AI training or accepting reduced visibility. The lower consistency of AIOs across identical queries and their fragility to minor edits highlight technical limitations that undermine reliability for critical information retrieval.

For the digital ecosystem, this research documents a significant shift in information gatekeeping power. Publishers face unprecedented visibility challenges, while users receive information filtered through proprietary models rather than diverse sources. The call for sustainable revenue frameworks acknowledges that current arrangements benefit AI providers at publishers' expense. This study provides concrete metrics showing how search market power concentration creates information asymmetries. The technical findings around source selection bias and consistency issues become increasingly important as generative search expands, potentially influencing content quality and diversity across the web.

• →AI Overviews appear in over 51% of queries and are displayed prominently above organic results, fundamentally changing search result presentation.
• →Generative search engines show significantly different source selection than Google Search, with preference for Google-owned content over institutional sources.
• →Websites blocking AI crawlers face systematic visibility disadvantages despite search engines having content access, creating problematic incentive structures.
• →AI Overviews demonstrate lower consistency across identical queries and reduced robustness to minor query variations compared to traditional search.
• →The study calls for new revenue and sustainability frameworks to create equitable ecosystems between publishers and generative search providers.