AI Search Engines Cite Less Popular Websites, New Research Shows
AI Search
Google
Gemini
Generative AI
Search Engines
Tranco
LLMs
Web Search
8
Shifting Sands
Media Hype
6/10
Real Impact
8/10
What is the Viqus Verdict?
We evaluate each news story based on its real impact versus its media hype to offer a clear and objective perspective.
AI Analysis:
While AI-powered search is generating considerable media attention, this research reveals a deeper, more nuanced reality about its sourcing habits, suggesting a longer-term shift in how information is accessed and utilized – a shift requiring careful monitoring and analysis.
Article Summary
A recent pre-print study from Ruhr University in Bochum and the Max Planck Institute for Software Systems has quantified the significant differences between traditional Google search results and those generated by AI-powered search engines. The research, comparing Google’s AI Overviews, Gemini-2.5-Flash, GPT-4o’s web search mode, and GPT-4o with Search Tool, found that generative search engines consistently cited websites significantly less popular than those appearing in the Top 10 of a traditional Google search. Researchers utilized test queries sourced from ChatGPT, AllSides, and Amazon’s most-searched products. Notably, AI engines favored corporate entities and encyclopedias, rarely citing social media, and often presented a similar level of conceptual coverage as traditional results. However, they also tended to compress information and, in some cases, omit nuanced details, particularly with ambiguous search terms. A key finding was that over 50% of AI Overviews’ sources didn’t appear in the Top 10 Google links, and 40% were outside the Top 100. While the study didn’t definitively declare AI search ‘better’ or ‘worse,’ it highlighted the need for new evaluation methods focusing on source diversity and synthesis behavior. This research has significant implications for understanding how AI is shaping the information landscape and how users should critically assess the results they receive.Key Points
- AI-powered search engines consistently cite less popular websites compared to traditional Google search results.
- The research highlights a difference in sourcing, with AI engines favoring corporate and encyclopedic sources.
- AI search engines often compress information and can omit nuanced details, particularly for ambiguous search terms.