Students Swap Google for Generative AI in Global Shift Toward Conversational Learning

Students are abandoning traditional search engines in favor of conversational AI, marking a fundamental shift in how the next generation accesses academic knowledge. A study led by researchers at Bar-Ilan University reveals that the habit of searching and comparing multiple sources is being replaced by direct interaction with tools like ChatGPT.

According to the research, which examined search activity across more than 20 countries between 2022 and 2025, there has been a global decline in Google searches and Wikipedia visits specifically for physics topics. The data indicates that this trend is not uniform; the drop in Google usage was significantly sharper in non-English-speaking countries than in the U.S., U.K., and Australia. Furthermore, the study found that AI adoption is highest for subjects that can be explained through text and logical reasoning. In contrast, topics requiring diagrams and graphs saw smaller declines in traditional search usage, as students still rely on visual data that AI often struggles to provide.

The End of the Discovery Era

For decades, online research followed a 'discovery' model. Students navigated various websites, evaluated the credibility of different sources, and synthesized fragmented information into a coherent understanding. This process served as a critical exercise in information literacy, forcing users to cross-reference facts and encounter diverse perspectives.

The rise of Large Language Models (LLMs) has introduced a 'consumption' paradigm. Instead of hunting for a source, a student can now use a single prompt to receive a direct, synthesized explanation. This transition lowers the barrier to entry for students who struggle with complex search queries or face language barriers when attempting to access scientific content predominantly published in English.

Implications for Global Education

This shift suggests a democratization of information, particularly for non-native English speakers who can now receive complex scientific concepts translated and simplified instantly. However, the reliance on text-based AI creates a potential vulnerability in learning. Because the decline in traditional search is less pronounced for visually complex subjects, there is a risk that students may over-rely on AI for logic-based tasks while neglecting the visual and spatial reasoning required for advanced science.

The Path Forward

As the academic world adjusts to this new behavior, the research—published in the journal Physical Review Physics Education Research—highlights a need for a systemic redesign of educational resources. Educators and publishers may need to develop tools that complement AI's ability to simplify text while reinforcing the critical thinking skills lost when students stop comparing sources.

What remains to be seen is whether this trend will extend beyond physics into other STEM fields and how the long-term erosion of 'search-and-compare' habits will affect the ability of students to verify the accuracy of AI-generated claims.