AI Chatbots and Knowledge Collapse are raising concerns about how students access and understand information. A University of Copenhagen-led study tested 27 large language models across 155 topics and found that chatbot responses were less diverse than information available through Google search. Researchers found that even GPT-5, the most diverse model tested, provided at least 18.7% less varied information than Google. They warn that repeated exposure to similar claims and perspectives could narrow the range of knowledge users encounter. The researchers describe a possible long-term risk as “knowledge collapse”, particularly if future AI models are increasingly trained on AI-generated content. However, the study does not conclude that knowledge collapse has already occurred. Students are encouraged to cross-check AI responses with textbooks, official sources, multiple websites and classroom discussions to develop broader understanding.
AI Chatbots and Knowledge Collapse: What Students Should Know
