A recent study analyzing 2.5 million full-text scientific papers from 2015 to 2025 found that large language models (LLMs) have not yet significantly transformed scientific practices. While LLMs emerged in late 2022, most pre-2023 trends in lexical, structural, epistemological, and social dimensions of science continued with limited detectable breaks. Notable exceptions included an acceleration in coauthor team growth within computer science in Asia and a perceived decline in biomedical topic diversity, which was attributed to standardization in article formats. The research provides an empirical baseline for tracking the health of global science as LLM adoption increases. AI
IMPACT This research provides an empirical baseline for understanding the actual impact of LLMs on scientific progress, rather than speculative claims.
RANK_REASON The item is an academic paper published on arXiv discussing research findings. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- arXiv
- CatalyzeX Code Finder for Papers
- computer science
- Connected Papers
- DagsHub
- Gotit.pub
- Hugging Face
- large language models
- Litmaps
- science
- ScienceCast
- scite Smart Citations
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →