Generative AI Increases Academic Preprint Productivity
Researchers from UC Berkeley and Cornell, publishing in Science, analysed more than one million preprints from 2018–2024 to study AI's effect on academic writing. They found authors' preprint output rose 36.2–59.8% after AI adoption (43–89.3% for Asian authors), and AI-generated complex language correlated with lower publication rates, suggesting linguistic masking of weak work. The study also found AI-enhanced Bing Chat recommended more diverse, recent sources, prompting calls for stronger methodological review and AI-assisted screening.
Key Points
- 1Measured authors' preprint output rose 36.2–59.8% after AI adoption, 43–89.3% for Asian authors
- 2Found AI-generated complex language correlated with lower publication rates, implying linguistic masking of weak science
- 3Recommend reviewers increase methodological scrutiny and adopt AI-assisted screening tools to detect low-quality submissions
Scoring Rationale
Strong peer-reviewed evidence of broad AI effects on publishing plus actionable implications; limitation is observational, not causal proof.
Sources
Public references used for this report.
Practice with real FinTech & Trading data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all FinTech & Trading problems
