• Source:JND
Add as a preferred

Chatbots and search engines may share the goal of delivering accurate answers, but a new study suggests they take entirely different routes to get there. While Google’s search relies on ranking and visibility, AI chatbots seem to venture much deeper into the internet — well beyond what most users ever see.

Researchers from Ruhr University Bochum and the Max Planck Institute for Software Systems explored how Google’s search engine compares to Google’s own AI Overviews, Gemini 2.5, and OpenAI’s GPT-4o. Their tests covered a variety of questions—from politics and factual topics to shopping recommendations—to see where each system drew its information from.

ALSO READ: PS6 First Details Revealed: Sony’s Next-Gen Console Could Redefine Gaming

Chatbots Go Deeper, Google Stays on Top

The researchers found that AI chatbots often gathered material from a far wider mix of websites than Google’s traditional search results. In fact, chatbots sometimes pulled data from pages buried beyond Google’s top 1,000 results — even from domains outside the top one million websites globally.

When it came to e-commerce-related queries, the overlap between Google’s results and those from chatbots was surprisingly low — under 30 percent. Even across all query categories, the similarity rarely exceeded 50 percent. Gemini stood out for pulling from lesser-known or lower-ranked websites.

Depth Over Popularity

While this might sound like chatbots are relying on obscure or unreliable sources, the study found otherwise. GPT-based systems often referenced formal, verified, or institutional websites — steering clear of social media content. Instead of starting from zero, these models used external data to validate or expand on their internal knowledge base.

Google, on the other hand, doesn’t operate with prior understanding. Its engine ranks pages based on factors like relevance, popularity, and search optimisation — aiming to serve what’s most visible and most likely to satisfy the majority of users.

Understanding the Shift

The study also observed that AI systems are more concerned with substance than presentation. They can parse complex research papers or long reports without simplifying them, prioritising reliability and detail over ranking position.

ALSO READ: Government Issues High-Severity Warning for Android Users: Update Your Phone Now

The researchers stopped short of declaring one method better than the other but highlighted a growing need to rethink how we evaluate information retrieval. As AI continues to shape how we search, the mystery of how chatbots choose and trust their sources is only deepening.

Also In News