Why Traditional Web Search Outperforms AI Learning Methods

Since ChatGPT’s 2022 launch, millions have turned to large language models (LLMs) for quick learning. However, emerging research indicates this convenience may hinder deeper understanding. Professors Jin Ho Yun and I conducted studies with over 10,000 participants, comparing learning through LLMs and traditional Google searches. Findings revealed that those using LLMs developed shallower knowledge, produced shorter and less informative advice, and engaged less with the material. Active engagement—navigating web links and synthesizing information—fosters deeper learning, while LLMs promote a passive approach. While LLMs offer undeniable benefits for quick answers, users must strategically leverage them for in-depth knowledge acquisition. Future research aims to explore generative AI tools that encourage active learning, especially crucial in education, where equipping students with foundational skills amid AI integration is essential. Understanding when to use LLMs versus traditional methods can enhance learning outcomes.

Source link