Scientific production in the era of large language models

2025/12/18 by Keigo Kusumegi, Xinyu Yang, Paul Ginsparg +4 · 4 voices · 55 citations
Business, Management and Accounting · Medicine · Social Sciences · #International Science and Diplomacy #Knowledge production #Process (computing) #Production (economics) #Production model #Science, Research, and Medicine #Scientific modelling #University-Industry-Government Innovation Models

paper · open access · doi:10.1126/science.adw3000

published in Science 390(6779), 1240-1243 (American Association for the Advancement of Science)

openalex created_date 2025/12/18 · openalex publication_date 2025/12/18 · openalex updated_date 2026/08/05

Abstract

With the production process rapidly evolving, science policy must consider how institutions could evolve.

Citations

• Quantifying large language model usage in scientific papers
• Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers
• Undermining Mental Proof: How AI Can Make Cooperation Harder by Making Thinking Easier
• Delving into LLM-assisted writing in biomedical publications through excess vocabulary
• Mathematical discoveries from program search with large language models
• Can large language models provide useful feedback on research papers? A large-scale empirical analysis
• Experimental evidence on the productivity effects of generative artificial intelligence
• Generative AI at Work
• The Virtual Lab of AI agents designs new SARS-CoV-2 nanobodies

Cited by

• Messy Research, Certification and the Monetization of Science
• The unintended consequences of large language models as a labor-augmenting technology in science
• Comment on Scientific production in the era of large language models
• LLM hallucinations in the wild: Large-scale evidence from non-existent citations
• How Open Must Language Models be to Enable Reliable Scientific Inference?
• How AI use in scholarly publishing threatens research integrity, lessens trust, and invites misinformation
• Benefits and risks of generative AI in scientific research: a two-layer analysis framework of biomedical publications
• The creeping normality of AI in the life sciences
• Intelligence Without Integrity: Why Capable LLMs May Undermine Reliability
• Generative AI and the future of scientometrics: current topics and future questions
• Safeguarding the role of humans in conservation science in the age of AI
• Generative Artificial Intelligence in Scientific Research: Individual Benefits, Collective Risks, and a Framework for Responsible Research with AI
• Academic journals' AI policies fail to curb the surge in AI-assisted academic writing
• Does Scientific Writing Converge to U.S. English? Evidence from Generative AI-Assisted Publications
• The Rise of Large Language Models and the Direction and Impact of US Federal Research Funding
• Visualizing the Adoption of Large Language Models across Sociology Subfields
• The Last Human-Written Paper: Agent-Native Research Artifacts
• The diffusion of large language models in published academic articles
• The next unit of science: Is the scientific paper due to be replaced?
• More Versus Better: Artificial Intelligence, Incentives, and the Emerging Crisis in Peer Review
• Scientific Writing, Generative Artificial Intelligence, and the Non–Native English Speaker
• Tracking claim changes from preprint to publication across 72,644 biomedical studies using large language models
• A robust association between LLM use and scientific productivity: Assessing stopping-time selection
• Beyond Via: Analysis and Estimation of the Impact of Large Language Models in Academic Papers
• Artificial Intelligence in Science: Returns, Reallocation, and Reorganization
• AI as a Partner in Learning about, Doing, and Engaging with Science: Vigilance as the Key to Productive Augmentation
• Benchmarking AI Performance on End-to-End Data Science Projects
• SciIntegrity-Bench: A Benchmark for Evaluating Academic Integrity in AI Scientist Systems
• Hitting a Moving Target: Test-Time Adaptation for AI Text Detection under Continual Distribution Shift
• Process-Oriented Evaluation of AI-Assisted Scientific Writing
• The Agentification of Scientific Research: A Physicist's Perspective
• Reimagining plant science training in the era of generative artificial intelligence: a global perspective. [europepmc]

Discussions

• Hurjaa miten nopeasti tekoälyt mullistavat myös tiedettä. Pari vuotta sitten vielä ihan tyhjänpäiväistä, ja nyt @science.org:ssa pohditaan, miten koko systeemi on mullistumassa - ja tässä vaiheessa ky [bsky, 13 points, 3 comments]
• The robots are not coming for your jobs, but bullshitters with robots ARE competing for your jobs, and unless we decouple publishing metrics from career paths (financial gain), we are going to lose bo [bsky, 3 points, 0 comments]
• Rachel Burley (Chief Publications Officer on scientific rigor in an age of LLMs): Surge after 2022 and release of ChatGPT. Points to paper which found researchers using AI tools wrote 36% more papers. [bsky, 2 points, 1 comments]
• #icanhazpdf anyone got this? With the production process rapidly evolving, science policy must consider how institutions could evolve doi.org/10.1126/scie... [bsky, 0 points, 0 comments]

Related

• Scientific production in the era of large language models
• The unintended consequences of large language models as a labor-augmenting technology in science
• Large language models have learned to use language
• How should the advancement of large language models affect the practice of science?
• How should the advent of large language models affect the practice of science?