DigestAI news desk

Cut through the AI noise.

Generative AI & Models1 min read

Study finds AI chatbots offer narrower knowledge than Google search

A University of Copenhagen study compared 27 large language models from OpenAI, Meta, Google, and Alibaba against Google search. Researchers tested answers on 155 topics across 12 countries, using 200 prompt variations per topic. The models generated about 1.7 million answers and 70 million claims, revealing that AI chatbots provided less varied responses than search engines.

1 source

Key points

  • University of Copenhagen study tested 27 AI models from OpenAI, Meta, Google, and Alibaba on 155 topics
  • GPT-5 showed **18.7% less variety** than Google search in answers across 12 countries
  • Smaller models and newer versions performed better than older ones but still lacked Google’s breadth

The study found that even OpenAI’s GPT-5, the most diverse model tested, delivered at least 18.7% fewer perspectives than Google. Smaller models and newer versions performed better than older ones, but none matched Google’s breadth. The issue stems from how AI models compress training data, prioritizing common patterns and filtering out outliers. Search engines, by contrast, expose users to multiple sources and viewpoints, broadening knowledge exposure.

The story so far

2 episodes →
  1. Study finds AI chatbots offer narrower knowledge than Google searchthis story
Full story from digitaltrends.com · by Manisha Priyadarshini · via Search: GoogleOpen source ↗

AI chatbots serve you a narrower slice of knowledge than a plain old Google search, study finds

digitaltrends.com · 28 September 2026

Loading the full article…

This text was published by digitaltrends.com and written by Manisha Priyadarshini. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗

Topics · follow one to build your own front page
University of CopenhagenOpenAIMetaGoogleAlibabaGPT-5

The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.

Comments

via GitHub Discussions

More in Generative AI & Models

All →

Related stories