From Search to GenAI Queries: Global Trends in Physics Information-Seeking Across Topics and Regions

The emergence of generative artificial intelligence (GenAI) marks a potential inflection point in the way academic information is accessed, raising fundamental questions about the evolving role of search in student learning. This study examines this shift by analyzing longitudinal trends in physics-related search and page-view activity, using declines in traditional search behavior as a quantitative proxy for changes in independent information-seeking practices. We analyze Google Trends data for core concepts in Classical Mechanics and Electromagnetism across three academic years (2022-2025) in more than 20 countries, and complement this analysis with Wikipedia page-view data across seven major languages to establish platform independence. The results reveal a substantial, systematic, and persistent global decline in search and page-view activity across most examined physics topics. The magnitude of this decline is domain-dependent, with Mechanics-related content exhibiting sharper and more consistent reductions than Electromagnetism-related content. Pronounced geographic and linguistic heterogeneity is observed: while English-speaking regions show relative stability or only moderate declines, non-English-speaking regions exhibit substantially larger reductions in traditional, search-based information-seeking activity. Despite the overall decrease in volume, the seasonal structure characteristic of academic activity remains robust. Taken together, these findings indicate a redistribution of physics-related information-seeking behavior in academic contexts where generative tools are increasingly available.

💡 Research Summary

This paper investigates how the rapid diffusion of generative artificial intelligence (GenAI) is reshaping the way physics students seek information online. Using two publicly available, platform‑independent data sources—Google Trends and Wikipedia page‑view statistics—the authors track longitudinal changes in search and viewing activity for core concepts in Classical Mechanics and Electromagnetism over three full academic years (August 2022 – August 2025).

Google Trends provides a Relative Search Volume (RSV) metric that normalizes the number of queries for a given topic against total search activity in a region and scales the peak to 100. The authors employ the “Science” category filter and Google’s “Topic” entities to aggregate synonyms, misspellings, and translations, thereby ensuring semantic consistency across languages while avoiding sparsity in low‑volume regions. A cross‑regional normalization further rescales each observation relative to the global maximum across all countries and weeks, mitigating biases from population size and internet penetration.

The temporal window is divided into three analytically defined epochs: Reference Year (2022‑2023), Transition Year (2023‑2024), and Integration Year (2024‑2025). Weekly RSV values (N = 53 per year) are aggregated into annual means (µ_y). The primary effect size is the percentage change %∆ = (µ_25 − µ_23)/µ_23 × 100. Statistical significance is assessed with paired‑sample t‑tests that match identical weeks across years, controlling for the strong academic seasonality (examination spikes, holiday troughs). Cohen’s d is reported as the practical significance metric, with conventional thresholds (0.2 small, 0.5 medium, 0.8 large).

Topic selection follows a systematic curation based on the table of contents of a standard introductory physics textbook. Only pure physics concepts are retained; auxiliary mathematical terms are excluded. Topics with prolonged zero values or erratic collapses are removed to ensure stable time‑series. The final set comprises roughly ten Mechanics concepts (e.g., Newton’s laws, kinetic energy, simple harmonic motion) and ten Electromagnetism concepts (e.g., electric field, magnetic flux, LC circuits).

Geographically, the study begins with the 40 most populous nations (population ≈ 38 million threshold) and narrows to a sample of over 20 countries where Google search data are reliable and sufficiently dense. Countries with insufficient signal (e.g., some large nations where physics‑related queries are sparse) are omitted, as is mainland China because Google is not the dominant search engine there. The sample intentionally includes a mix of English‑speaking (US, UK, Australia), Romance‑language (Spain, France, Mexico), and Arabic‑language (Saudi Arabia, Egypt) nations to capture linguistic heterogeneity.

Key findings:

1. Global Decline – Across the board, physics‑related RSV and Wikipedia page‑views have dropped relative to the 2023 baseline. Mean percentage changes range from –12 % to –28 % depending on topic and region.

2. Domain‑Specific Patterns – Mechanics topics experience steeper declines (average ≈ –25 %) than Electromagnetism topics (average ≈ –15 %). The authors suggest that Mechanics, being more concrete and traditionally taught via textbook problem sets, may be more readily supplanted by AI‑generated explanations.

3. Geographic & Linguistic Heterogeneity – English‑dominant regions show modest reductions (‑5 % to ‑10 %). Non‑English regions (Spanish, French, Arabic) exhibit markedly larger drops (‑20 % to ‑35 %). This disparity aligns with prior surveys indicating variable GenAI adoption rates, trust levels, and infrastructure across countries.

4. Persistence of Seasonal Structure – Despite overall volume loss, the classic academic rhythm (exam‑time spikes, holiday troughs) remains evident in the time series, confirming that students continue to align information‑seeking with curricular milestones.

The authors interpret the systematic decline in traditional search activity as a proxy for a shift toward GenAI‑mediated information retrieval. They argue that the observed patterns likely reflect students increasingly relying on conversational AI tools (e.g., ChatGPT, Gemini) for quick answers, conceptual explanations, and problem‑solving assistance, thereby reducing the need to formulate explicit search queries.

Limitations are acknowledged: the study does not directly measure GenAI usage; causality between search decline and AI adoption remains inferential. Moreover, reliance on Google and Wikipedia excludes offline or alternative platform behaviors. The authors call for future work that integrates AI interaction logs, learning management system data, and performance outcomes to more precisely map the causal chain from AI tool adoption to learning behavior and achievement.

In conclusion, the paper provides robust, cross‑regional evidence that traditional web‑search–based physics information‑seeking is waning, especially in non‑English contexts, while the academic seasonal cadence persists. These trends suggest a substantive reconfiguration of the digital learning ecosystem driven by generative AI, with implications for curriculum design, digital literacy initiatives, and institutional support for responsible AI integration.